Can Environmental Factors Predict Precious Metal Returns?

Can Environmental Factors Predict Precious Metal Returns? | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Can Environmental Factors Predict Precious Metal Returns? Ilyes Abidi, Kamel Touhami, Mariem Nsaibi, Maissa Mejri This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6099418/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study examines the influence of environmental factors, including CO₂ emissions and temperature anomalies, on the returns of precious metals (gold, silver, platinum) using DCC-GARCH and TVC-VAR models. The results show that gold exhibits relative stability with predictable returns, while silver and platinum, being more volatile, are subject to more pronounced fluctuations. CO₂ emissions and climate anomalies influence the volatility of precious metals by disrupting their supply and demand, while amplifying macroeconomic uncertainties and production costs, leading to increased persistence of past shocks on current volatility. The ARCH and GARCH coefficients reveal a strong persistence of these shocks for all metals, which is crucial for long-term forecasting. The dynamic conditional correlations captured by the DCC-GARCH model highlight critical interdependencies between the returns of precious metals and environmental variables. The accumulated impulse response functions show varying adjustments of precious metals to environmental shocks, reflecting a gradual adaptation of markets to new climate realities. Forecasts from the TVC-VAR model confirm the relevance of gold as a safe-haven asset, while silver and platinum require proactive management to mitigate the risks associated with their dual roles as both industrial and precious metals. These findings encourage investors, policymakers, and businesses to adopt sustainable and proactive strategies in the face of environmental challenges, while enriching the understanding of the complex interactions between finance and climate. Business and commerce/Economics Business and commerce/Finance Social science/Environmental studies CO₂ temperature anomalies gold silver platinum DCC-GARCH TVC-VAR Figures Figure 1 1. Introduction In an ever-evolving global economic context marked by geopolitical crises, financial tensions, and a rapid transition to sustainable economic models, precious metals play a key role in investors' portfolios looking to secure their wealth. Historically seen as safe havens, these assets continue to be valued for their ability to preserve wealth in times of economic uncertainty, particularly amid financial market volatility, persistent inflation, and geopolitical disruptions. In 2022, gold reached a record global demand, approaching 4,700 tons, highlighting its essential role in risk management (World Gold Council, 2022 ). However, beyond traditional economic factors, environmental elements are increasingly influencing the dynamics of precious metal markets. The extraction of precious metals, while fundamental to global supply, presents significant environmental challenges. For instance, producing just one gram of gold can generate up to 20 tons of mining waste, resulting in considerable CO2 emissions and damage to local biodiversity (ICMM, 2019). In response to these concerns, stricter regulations have been imposed, increasing extraction costs for some gold mines by up to 30% (World Gold Council, 2020 ). Consequently, there has been a surge in demand for investments in responsible mining companies, with more than $ 50 billion in assets invested in 2020 (BlackRock, 2021 ). These developments illustrate the growing pressure for more sustainable production, supported by initiatives like the "Responsible Gold Mining Principles" from the Council on Mining and Metals (ICMM). Alongside this trend, the transition to a green economy is reshaping the demand for these resources. Precious metals, seen as sustainable materials, now play a central role in discussions about energy transition and the environmental impact of mining. Economic actors must rethink their understanding of market dynamics, incorporating factors such as climate risks, ESG (Environmental, Social, and Governance) policies, and technological innovations in renewable energy. Previous research on the predictability of precious metal returns has mainly focused on traditional economic factors such as interest rates, inflation, and geopolitical shocks. Urquhart ( 2017 ) highlighted the complexity of short-term prediction of these assets, often regarded as safe havens, due to the unpredictable volatility of markets. In this line, Dichtl ( 2020 ) demonstrated that incorporating long-term financial indicators improves the robustness of forecasts, though they remain limited by the high sensitivity of precious metals to financial market fluctuations. Taking a broader perspective, Papenfuß et al. ( 2025 ) analyzed the evolution of predictive factors for precious metals and found that the influence of autoregressive price components has gradually diminished, with financial market indices becoming more significant, particularly after the 2008 crisis. In light of the limitations of traditional approaches, Shahid et al. ( 2020 ) applied the AMH (Adaptive Market Hypothesis) theory to precious metals and demonstrated the temporal variability in their predictability, with investors continuously adjusting their strategies based on new economic and environmental information. This dynamic is particularly pronounced during crises, making classical models less effective. In this context, Zhang and Pan ( 2021 ) highlighted the impact of economic and geopolitical crises on the volatility of precious metal prices, especially in relation to oil market fluctuations. Similarly, Huang et al. ( 2022 ) confirmed gold's safe-haven role, while tempering this function for silver and platinum, which are more exposed to industrial demand fluctuations. The importance of economic uncertainties was also highlighted by Raza et al. ( 2023 ) through the GARCH-MIDAS approach, which integrates global macroeconomic variables and improves forecast accuracy. At the same time, Yang et al. ( 2024 ) showed that while gold and oil are influential assets, they do not serve as safe havens for ESG assets but contribute to portfolio diversification. In a complementary approach, Karmakar et al. ( 2023 ) used an INGARCH model to analyze the impact of climate risks on trading volumes of gold futures, revealing that physical risks significantly influence these volumes in the short term. These advances illustrate that the predictability of precious metal returns is constantly evolving, influenced by structural changes and new investment paradigms. The integration of environmental risks and adaptive approaches is therefore crucial for refining predictive models. However, these models still face major challenges, particularly with climate shocks that directly disrupt precious metal production. For example, in 2021, extreme weather events reduced gold production by 15% in certain regions (Metals Focus, 2021 ), highlighting the industry's vulnerability to environmental disruptions. These factors emphasize the need to integrate climate risks more thoroughly into analyses to better anticipate long-term impacts on the precious metal market. At the same time, the demand for precious metals used in green technologies, such as electric vehicles and solar panels, is growing rapidly. According to Research and Markets ( 2022 ), the electric vehicle market, reliant on palladium, platinum, and silver, is expected to reach a value of $ 800 billion by 2027. These transformations in the demand and supply of precious metals call for a revision of traditional forecasting models, which struggle to integrate these new dynamics. This study aims to examine the impact of environmental factors on the predictability of precious metal returns, specifically focusing on gold, silver, and platinum, as well as environmental variables such as CO2 emissions and temperature anomalies. The main objective is to analyze how these environmental variables influence price volatility and evaluate how markets respond to risks associated with the energy transition, while providing methodological tools to help investors better integrate these environmental factors into their investment decisions. The methodological framework relies on two advanced econometric models: the DCC-GARCH (Dynamic Conditional Correlation Generalized Autoregressive Conditional Heteroskedasticity) model and the TCV-VAR (Time-Varying Coefficients Vector Autoregression) model. The DCC-GARCH model is used to analyze the dynamic conditional correlations between precious metal returns and environmental factors, offering a detailed understanding of market adjustments to ecological changes and energy transition policies. The TCV-VAR model, on the other hand, allows for the examination of the dynamic interdependencies between these variables and captures the time-varying nature of relationships between precious metal returns and environmental variables by modeling nonlinear and time-changing effects. This ability to adapt coefficients over time is crucial for capturing the short- and long-term impacts of environmental shocks. Both models are estimated using daily data covering the period from January 5, 2017, to October 2, 2023, allowing for an in-depth analysis of temporal dependencies and conditional heteroskedasticity, essential for obtaining reliable results in this study context. While classical econometric models like GARCH-MIDAS, applied by Nguyen and Walther ( 2020 ) and Dinh et al. ( 2022 ), have evaluated the long-term impact of macroeconomic factors on the volatility of precious metals, these approaches have notable limitations. The GARCH-MIDAS mainly focuses on integrating long-term macroeconomic variables without considering structural changes in coefficients over time. Furthermore, AI-based approaches, such as those proposed by Pierdzioch and Risse ( 2020 ) and Cohen ( 2022 ), have indeed improved the prediction of returns but suffer from a lack of economic interpretability, making it difficult to understand the underlying mechanisms between financial markets and environmental variables. The use of alternative data, such as internet search trends explored by Miao et al. ( 2022 ), remains limited as it cannot effectively assess interactions with specific environmental indicators. In response to these gaps, our research stands out by explicitly integrating environmental factors into the predictability of precious metal returns. The TCV-VAR model allows for analyzing the dynamic evolution of relationships between precious metal returns and environmental variables, capturing nonlinear and time-varying effects. Unlike classical models, it incorporates the adaptability of coefficients over time, providing a better understanding of the short- and long-term impacts of environmental shocks. Additionally, the DCC-GARCH model is used to evaluate the dynamic conditional correlations between precious metal returns and environmental indicators, offering a nuanced analysis of market adjustments to ecological changes and energy transition policies. By combining these two approaches, our study fills a major methodological gap in the existing literature. It provides a more comprehensive and nuanced perspective on the influence of environmental factors on the dynamics of precious metals while meeting the needs of investors, regulators, and policymakers in the context of the transition to a greener economy. This paper is structured as follows. Section 2 offers a review of relevant literature. Section 3 details the methodological framework and data analysis. The main results are discussed in Section 4 . Finally, Section 5 presents the conclusions and implications of this study. 2. Literature Review Precious metals, due to their historical role as a store of value, continue to attract the attention of investors and researchers. The predictability of their returns has been the subject of numerous studies, particularly during periods of high financial market volatility. While traditional economic factors, such as inflation, interest rates, and financial crises, have long been considered the primary determinants of these returns, more recent research has highlighted the growing importance of environmental, social, and governance (ESG) variables. At the same time, the rise of advanced quantitative methods and artificial intelligence technologies has significantly improved the forecasting capabilities of existing models. These developments offer a better understanding of the mechanisms influencing precious metal prices and open new perspectives for analyzing their behavior in financial markets. Historically, the returns of precious metals have been heavily influenced by macroeconomic variables and global economic cycles. Urquhart ( 2017 ) analyzes the predictability of precious metal returns through classical econometric models (Variance ratio test, Brock, Dechert, and Scheinkman test, Hurst exponent) and emphasizes that, although these assets are often considered safe havens, their short-term predictability remains complex. Price fluctuations can be exacerbated by unforeseen events, such as financial crises, making predictions uncertain. In this regard, Dichtl ( 2020 ) explores the possibility of using methodologies from stock markets to forecast excess returns of gold. He shows that models incorporating long-term financial indicators provide more robust forecasts, but these remain limited by gold's high sensitivity to financial market dynamics. Thus, these studies highlight that the stability of financial markets plays a determining role in the predictability of precious metals, but that this predictability is conditioned by the ability of models to integrate often unpredictable external factors. In this context, Papenfuß et al. ( 2025 ) examine the evolution of predictive factors for metal prices in a context marked by phases of financialization and de-financialization of commodity markets. Through the analysis of 24 metals over the period 1995–2019, they demonstrate that the autoregressive components of prices are the primary determinants, although their influence gradually decreases. They also highlight the importance of interest rates before the 2008 financial crisis, while financial market indices play a more dominant role afterward. In terms of predictive performance, their models significantly outperform traditional benchmarks, such as the random-walk model, in 12 out of the 24 cases studied, particularly for minor metals. These results confirm that the relationships between macroeconomic variables and metal prices are dynamic and evolve over time, justifying the use of flexible forecasting models adapted to the structural changes of the market. In the face of these challenges, some researchers have adopted an alternative approach based on the Adaptive Market Hypothesis (AMH), which considers that investors adjust their strategies according to new economic and environmental information. Shahid et al. ( 2020 ) apply this approach to precious metals and show that their returns evolve according to an adaptive process, reflecting investors' constant adjustments in response to economic fluctuations and market shocks. This theoretical framework helps explain why return predictability is not a fixed characteristic but varies over time depending on macroeconomic conditions and market expectations. This adaptability is especially pronounced in times of crisis, where uncertainty amplifies investors' adjustments, making traditional models less effective. Another key aspect of this predictability is the impact of economic and geopolitical crises on precious metals. Zhang and Pan ( 2021 ) show that the price volatility of gold and platinum is significantly influenced by fluctuations in the oil market, particularly during periods of global economic tension. Financial crises, such as the 2008 crisis or the COVID-19 pandemic, have strengthened the appeal of precious metals as safe-haven assets, leading to an increase in their returns. Huang et al. ( 2022 ) confirm this trend by analyzing the ability of precious metals to serve as a hedge against geopolitical and economic risks. However, they note that this safe-haven function varies across metals, with gold being historically more resilient than silver or platinum, which remain more susceptible to industrial fluctuations. In this perspective, understanding the dynamics of precious metal volatility during periods of economic uncertainty becomes essential. Modeling this volatility is thus a central issue, allowing for price movements to be anticipated in response to macroeconomic shocks. Raza et al. ( 2023 ) apply the GARCH-MIDAS approach to examine the impact of global economic uncertainties on precious metal volatility, particularly during the COVID-19 pandemic. Their findings show that incorporating global macroeconomic variables improves forecasting accuracy, highlighting the sensitivity of precious metals to economic uncertainty shocks. While precious metals have long been favored for their role as protection against economic crises, more recent financial instruments, such as green bonds, have begun to emerge as viable alternatives. Huang et al. ( 2022 ) analyze the effectiveness of green bonds as a hedge against economic shocks and conclude that they offer more sustainable stability in the face of environmental and social risks. Unlike precious metals, which strongly react to economic cycles and financial crises, green bonds allow for more effective diversification in long-term investment contexts. These findings open new perspectives for investors seeking to integrate ESG criteria into their strategies, while questioning the role of precious metals in a changing financial environment. In this regard, Yang et al. ( 2024 ) examine the predictability of ESG stock returns, taking into account the role of traditional assets like gold and oil, as well as uncertainties related to the market, cryptocurrencies, and geopolitical risks. By applying non-parametric techniques such as quantile causality and quantile-on-quantile regression, their results show that gold, oil, market-implied volatility (VIX and OVX), and geopolitical risk are significant predictors of ESG stock returns. However, neither gold nor oil play the role of a safe haven for these assets but rather serve to diversify portfolios. On the other hand, ESG stocks appear to be an effective hedge against geopolitical shocks and uncertainties related to cryptocurrencies during bearish market periods. These findings underscore the importance of dynamic portfolio management for sustainable investments and strengthen the literature on the resilience of ESG stocks in the face of financial and environmental crises, while guiding investors on the strategic implications of these instruments in asset allocation and risk management. Moreover, de Karmakar et al. ( 2023 ) explore the impact of both physical and transition climate risks on the trading volume of gold futures contracts. Due to the nature of the data, which is count-type, the authors use an INGARCH (log-linear Poisson GARCH-type) model to predict these volumes based on covariates related to climate risks. Their analysis reveals that physical risks have significant predictive power for gold trading volumes at 5-day and 22-day horizons. Additionally, a positive relationship between physical risks and gold trading volumes is observed, indicating that gold serves as a hedge against short-term physical risks (1 week and 1 month). Similar results are found for platinum and palladium, but not for silver. The authors emphasize the importance of gold as a safe-haven asset in the face of climate risks, especially physical risks, and demonstrate that this analysis provides a first direct approach to predicting the trading volumes of precious metals using count-based data models. At the same time, technological advancements have significantly transformed the methods for forecasting precious metal returns. Pierdzioch and Risse ( 2020 ) explore the use of random forests, a supervised learning technique, to predict the returns of precious metals. Their results show that these models capture complex relationships between various economic variables and the prices of precious metals, offering an alternative to traditional econometric models. In a similar approach, Cohen ( 2022 ) introduces advanced algorithmic strategies, based on artificial intelligence, to identify hidden patterns in the time series of precious metal prices. These emerging approaches pave the way for a significant improvement in forecasting capabilities, especially by incorporating external factors such as environmental and social data. Along the same lines, Mehrdoust and Noorani ( 2024 ) introduce a forecasting model based on neural networks optimized by a Lévy flight algorithm, enabling better handling of the complex price movements of precious metals. The results indicate that the neural network optimized by the Lévy flight algorithm outperforms other prediction models in terms of accuracy. This model offers a unique approach to predicting precious metal prices and can be applied to different time series. Furthermore, another promising approach involves leveraging online search trends to anticipate fluctuations in precious metal prices. Miao et al. ( 2022 ) study the impact of online search trends and show that these non-financial data can provide early signals about market developments. Using a non-parametric causality approach, they demonstrate that investor interest in specific keywords may be correlated with future price changes of precious metals. This new source of data, combined with machine learning models, could enhance investors' ability to anticipate market fluctuations. Finally, the non-stationary volatility of precious metals remains a challenge for long-term forecasting. Addison and Ghoshray ( 2023 ) explore this issue by applying models that incorporate changes in the variance of returns. Their results show that, while precious metals are often considered stable in the long term, they are subject to unpredictable variations, requiring adaptive models capable of absorbing these structural changes. Although existing studies have provided significant insights into the predictability of precious metal returns, several limitations remain. On the one hand, many studies focus primarily on traditional macroeconomic factors such as inflation, interest rates, and financial market volatility, without giving sufficient attention to environmental factors. For example, the GARCH-MIDAS and DCC-GARCH models applied by Nguyen and Walther ( 2020 ) and Dinh et al. ( 2022 ) have assessed the impact of long-term economic variables on precious metal volatility, but these models do not explicitly account for environmental factors. The GARCH-MIDAS, for instance, mainly focuses on integrating long-term macroeconomic factors but does not consider structural changes in coefficients over time. The same is true for the DCC-GARCH, which examines the dynamics of correlations between financial assets but does not adequately capture the dynamic impacts of environmental shocks or ecological policies on precious metal returns. On the other hand, the growing use of artificial intelligence and machine learning methods, as proposed by Pierdzioch and Risse ( 2020 ) or Cohen ( 2022 ), offers a notable improvement in forecasting returns. However, these approaches often suffer from a lack of interpretability and economic contextualization, limiting their applicability in understanding the underlying relationships between financial markets and environmental variables. Additionally, the study by Miao et al. ( 2022 ) on the use of online search trends to anticipate precious metal price variations highlights the importance of alternative data but does not assess their interaction with specific environmental indicators. In response to these limitations, our research stands out by explicitly integrating the role of environmental factors in the predictability of precious metal returns, using a combined approach of the Time-Varying Coefficients Vector Autoregression (TCV-VAR) model and the DCC-GARCH. The TCV-VAR model will allow us to analyze the dynamic evolution of relationships between precious metal returns and environmental variables, capturing non-linear and time-varying effects. Unlike classical models such as GARCH-MIDAS or DCC-GARCH, the TCV-VAR model allows for the modeling of the adaptability of variable coefficients over time, capturing both short-term and long-term effects of environmental shocks. Furthermore, the application of DCC-GARCH will enable us to assess the dynamics of conditional correlations between these returns and environmental indicators, providing a more nuanced understanding of how markets adjust to ecological changes and energy transition policies. By combining these two approaches, our study makes a significant contribution by bridging a methodological and conceptual gap in the existing literature. It enhances our understanding of how environmental factors influence the dynamics of precious metal returns and provides new tools for predicting price fluctuations in a rapidly evolving economic and ecological context. 3. Data and Methodology 3.1. Data This study aims to examine the impact of environmental factors on the predictability of precious metal returns, focusing specifically on gold, silver, and platinum, as well as environmental variables such as CO₂ emissions and temperature anomalies. The primary objective is to analyze how these environmental variables influence the volatility of precious metal prices and assess how markets respond to the risks associated with the energy transition. Additionally, the study provides methodological tools to assist investors in better integrating these environmental factors into their investment decisions. To address these questions, the empirical analysis relies on daily return data for three precious metals: gold, silver, and platinum, alongside daily data on carbon emissions and temperature anomalies. The return data for the precious metals were sourced from the Macrotrends website, a reputable provider of historical financial information. Regarding the environmental variables, carbon emissions were extracted from the Global Monitoring Laboratory, while temperature anomalies were retrieved from the Climate Reanalyzer platform. The data covers the period from January 5, 2017, to October 2, 2023. The returns for the precious metals are calculated from daily closing prices using the standard logarithmic return formula: R t =ln ( \(\:\:\frac{\text{P}\text{t}}{\text{P}\text{t}-1}\:\) ) Where P t and P t−1 represent the closing prices on consecutive days. Carbon emissions and temperature anomalies are also expressed in terms of daily values, using them directly as provided by the aforementioned sources. 3.2. Methodology To explore the impact of environmental factors on the predictability of precious metal returns, two econometric models are applied in this study: the DCC-GARCH (Dynamic Conditional Correlation Generalized Autoregressive Conditional Heteroskedasticity) model and the TCV-VAR (Time-Varying Coefficients Vector Autoregression) model. The DCC-GARCH model, introduced by Engle ( 2002 ), is used to analyze the dynamics of the conditional correlation between the returns of precious metals and environmental variables (carbon emissions and temperature anomalies). This model captures the effects of conditional volatility and changes in correlation relationships over time, which is particularly relevant for financial time series. The model is estimated in two steps: first, univariate GARCH models are fitted to the return series, and then conditional correlations are estimated, accounting for volatility. The formula for the DCC-GARCH model can be described as follows: First step (univariate GARCH models): Y t = µ t + ϵ t​ Where y t ​ is the return, µ t is the conditional mean, and ϵ t is the conditional error. Second step (estimation of conditional correlations): Σ t = Q + Aϵ t−1 ϵ′ t−1 A′ + BΣ t−1 B′ Where Σ t ​ is the conditional variance-covariance matrix, Q is the constant matrix, and A and B are matrices that determine the volatility dynamics. The TCV-VAR model, formalized by Swamy et al. ( 2010 ), is used to examine the dynamic interdependence between precious metal returns and environmental factors, while allowing the model's coefficients to vary over time. This enables the analysis of how the impact of environmental variables on precious metal returns evolves over different periods. The model is specified as follows: y t = C t + \(\:\sum\:_{i=1}^{p}\text{A}\) i y t−i + ϵ t Where y t ​is the vector of precious metal returns and environmental variables, C t is the time-varying constant vector, A i represents the matrices of time-varying coefficients, and ϵ t is the error vector. Both models are estimated using daily data and account for temporal dependencies and conditional heteroscedasticity effects, which is essential for obtaining reliable results in the context of a financial analysis of precious metals influenced by environmental factors. 4. Empirical Results and Discussion Table 1. Descriptive Statistics of Returns Variable Gold Silver Platinum CO2 Anomalies Mean 0.025 0.013 -0.003 0.000029 0.334 Maximum 6.789 8.896 11.176 0.000743 0.753 Minimum -5.400 -12.345 -12.315 -0.000956 -0.832 Standard Deviation 0.868 1.799 1.762 0.000312 0.224 Skewness -0.169 -0.495 -0.277 -0.721 -1.717 Kurtosis 5.399 6.779 4.536 0.801 4.224 Jarque-Bera 2116.160 3392.775 1510.625 197.25 2141.759 Probability 0.000 0.000 0.000 0.000 0.000 Observations 1729 1729 1729 1729 1729 Q² (10) 160.834 (0.000) * 284.542 (0.000) * 451.494 (0.000) * 5669.857 (0.000) * 7769.997 (0.000) * ARCH (10) 116.05 (0.000) * 155.06 (0.000) * 249.4 (0.000) * 1693.4 (0.000) * 1486.6 (0.000) * (*): indicates significance at the 5% level; Q²(10) : are the statistics from the Ljung-Box test with 10 lags applied to squared returns; ARCH (10) : is the heteroscedasticity test by Engle (1982). Table 1 presents the descriptive statistics of the returns for gold, silver, platinum, and environmental variables such as CO₂ emissions and temperature anomalies. The average returns reveal a slightly positive trend for gold (0.025) and silver (0.013), indicating a moderate appreciation of these precious metals over the studied period, while platinum shows a negative average (-0.003), suggesting a slight depreciation. These results are of great importance for investors wishing to diversify their portfolios with safe-haven assets. They align with numerous previous studies that indicate gold and silver often generate positive returns during periods of economic uncertainty. For example, the study by Baur and Lucey (2010) concludes that gold is a safe-haven asset, particularly during crises. However, our research stands out by including environmental factors, which seem to exert a more pronounced influence on the dynamics of precious metal returns, especially for platinum, which is more sensitive to industrial factors. As for volatility, the high standard deviations for silver (1.799) and platinum (1.762) compared to gold (0.868) highlight greater variability in returns, implying increased risk but also a higher potential for return, thereby attracting investors willing to tolerate higher volatility for potentially larger gains. The environmental variables, with much lower standard deviations (CO2: 0.000312; temperature anomalies: 0.224), show relative stability, reflecting the less volatile nature of these factors compared to precious metal returns. The negative skewness coefficients observed for gold (-0.169), silver (-0.495), platinum (-0.277), and environmental variables (CO2: -0.721; temperature anomalies: -1.717) indicate a left-skewed distribution of returns, meaning that significant losses are more frequent than substantial gains. This represents a significant downside risk for investors. This dynamic aligns with the work of Zhang and Pan (2021), who highlight that risk aversion influences return predictability, especially in the face of significant loss risks. Indeed, faced with negative skewness, investors are more likely to react negatively to the prospect of substantial losses, which can affect their perception of future returns and investment strategies. The interaction between this return skewness and risk aversion becomes even more critical in the context of precious metals, where environmental factors, such as CO₂ emissions and temperature anomalies, introduce additional uncertainties that may increase volatility and make returns even more unpredictable. Moreover, the high kurtosis observed for gold (5.399) and silver (6.779) reveals a leptokurtic distribution, characterized by thick tails and an increased probability of extreme events. This feature emphasizes the importance for investors to protect themselves against extreme risks, using appropriate financial instruments or further diversifying their portfolios. This finding echoes that of Dinh et al. (2022), who show that stock market volatility can also induce extreme events in metal markets. Thus, the presence of high kurtosis in precious metal returns reinforces the idea that investors must not only account for daily fluctuations but also the risks linked to more severe and less frequent environmental shocks. The results of the Jarque-Bera normality tests, all significant (p-value = 0.000), confirm that the returns for precious metals and environmental variables do not follow a normal distribution. This deviation from normality suggests that traditional approaches based on the normality assumption may be inadequate for modeling and forecasting returns. In line with Dichtl (2020), who emphasizes that forecasting techniques must account for these deviations to be effective, it becomes essential to adopt advanced models capable of better capturing the underlying dynamics of returns. Therefore, the use of models such as DCC-GARCH and TVC-VAR is particularly relevant to capture the complexity of interactions between precious metal returns and environmental factors. The Q² (10) and ARCH (10) statistics, both significant (p-value = 0.000), reveal the presence of autocorrelations and heteroscedasticity in the squared returns. This suggests the existence of volatility clusters, where periods of high volatility are followed by similar periods, a phenomenon typically observed in financial markets. For investors, this implies that periods of increased risk can be anticipated and managed more effectively using appropriate econometric models, thereby improving risk management and strategic decision-making. By integrating these analyses, it becomes clear that the returns of precious metals are influenced not only by traditional economic factors but also by environmental factors. Fluctuations in CO₂ emissions and temperature anomalies can affect the production and demand for precious metals, thus impacting their prices and returns. This interconnection highlights the importance for investors and policymakers to adopt a holistic approach in evaluating risks and opportunities in precious metal markets. In particular, considering environmental risks in investment strategies can lead to better portfolio diversification and reduced systematic risks associated with climate change and environmental regulations. These results are in line with the work of Huang et al. (2022), who emphasize the importance of external factors in the predictability of precious metal returns. However, our study deepens this perspective by more explicitly integrating environmental variables, thus highlighting their determining role in the evolution of precious metal markets. Table 2. Unit Root Tests for Returns ADF Stationarity Test PP Stationarity Test Variable No Constant With Constant With Constant and Trend No Constant With Constant With Constant and Trend Gold -40.756 (0.000) *** -40.778 (0.000) *** -40.776 (0.000) *** -40.841 (0.000) *** -40.908 (0.000) *** -40.912 (0.000) *** Silver -42.409 (0.001) *** -42.399 (0.000) *** -42.488 (0.000) *** -42.412 (0.001) *** -42.402 (0.000) *** -42.391 (0.000) *** Platinum -41.344 (0.000) *** -41.332 (0.000) *** -41.320 (0.000) *** -41.864 (0.000) *** -41.851 (0.000) *** -41.837 (0.000) *** Co2 -2.922 (0.003) *** -2.933 (0.041) *** -2.932 (0.152) -8.079 (0.000) *** -8.132 (0.000) *** -8.142 (0.000) *** Anomalies -1.558 (0.112) -4.416 (0.000) *** -4.407 (0.002) *** -2.130 (0.031) *** -5.256 (0.000) *** -5.243 (0.000) *** Note: The significance levels of 1%, 5%, and 10% are shown by ***, **, and *, respectively. Table 2 presents the results of the unit root tests, specifically the Augmented Dickey-Fuller (ADF) test and the Phillips-Perron (PP) test, applied to the returns of gold, silver, platinum, as well as the two environmental variables: CO2 emissions and temperature anomalies. The results of the ADF and PP tests for the returns of the three precious metals show negative and statistically significant test statistics at the 1% level (p < 0.01), indicating that these series are stationary, regardless of whether a constant or trend is included. For example, for gold, the ADF values range from -40.756 to -40.776, and the PP values range from -40.841 to -40.912, all significant at the 1% level. This means that the returns of gold, silver, and platinum do not contain a unit root, implying that shocks affecting these series are temporary and do not persist indefinitely. This is a crucial factor for investors looking to model the volatility and future returns of these assets. From a financial perspective, this suggests that the returns of these precious metals revert quickly to their mean after a shock, providing some predictability in their behavior. Regarding the environmental variables, the results are more mixed. For CO2 emissions, although the PP tests indicate strong stationarity with negative values (ranging from -8.079 to -8.142, all significant at the 1% level), the ADF test results are more nuanced. Without trend or constant, CO2 shows an ADF statistic of -2.922, significant at the 1% level, suggesting stationarity. However, when a constant and trend are included, the significance slightly decreases (p = 0.152), which could indicate the series' sensitivity to these components, potentially reflecting underlying trends not captured by the simple model. This highlights the importance of using a dynamic model to avoid incorrect conclusions. Regarding temperature anomalies, establishing stationarity is more complex. The ADF results without trend or constant (-1.558) are not significant, suggesting that the series could be non-stationary. However, with the addition of a constant and trend, the ADF values (-4.416 and -4.407) become significant at the 1% level, indicating conditional stationarity. The PP test results reinforce this conclusion, with statistics ranging from -2.130 to -5.256, all significant except without constant or trend. This suggests that temperature anomalies may have a trend component or other complex dynamics that require more sophisticated modeling to fully capture their behavior. From a financial perspective, this could suggest that returns linked to temperature anomalies are influenced by long-term trends or exogenous factors, making their forecasting more difficult and requiring a more rigorous approach to integrate them into forecasting models. In conclusion, the results of the ADF and PP tests show that the calculated test statistics for all series are well below the critical values at the 1%, 5%, and 10% significance levels for the three models (without constant or trend, with constant, and with constant and trend). It is noteworthy that the CO2 emissions series is integrated of order 1. These results allow us to reject the null hypothesis of the presence of a unit root for all series, confirming the stationarity of the studied series. Table 3: Estimations of the DCC Model (1.1) (Gold, CO2, and Temperature Anomaly) Gold Co2 Anomalies Panel A : Mean Equation Constant 0.010 (0.567) 0.000 (22.408) * 0.386 (56.886) * Panel B : Variance Equation Constant 0.009 (1.314) 0 (0.000) 0.001 (9.622) * ARCH(α) 0.051 (2.462) * 0.170 (12.162) * 0.975 (21.722) * GARCH(β) 0.937 (34.606) * 0.828 (64.177) * 0 (0.00) α+β 0.988 0.998 0.976 Panel C : Dynamic Conditional Correlation A 0.117 (7.534) * B 0.744 (18.866) * Log-Likelihood 11628.19 Panel D : Residual Diagnostics Test Q (20) 36.433 (0.01) 22039 (0.000) 22110 (0.000) (*) Indicates statistical significance at the 5% level; Q (20) represents the autocorrelation test statistics applied to the residuals. Table 4: Estimation of the DCC-GARCH Model (1.1) (Silver, CO2, and Temperature Anomaly) Silver Co2 Anomalies Panel A : Mean Equation Constant -0.017 (-0.499) 0.000 (22.408) * 0.386 (56.886) * Panel B : Variance Equation Constant 0.017 (1.900) 0 (0.000) 0.001 (9.622) * ARCH(α) 0.038 (5.662) * 0.170 (12.162) * 0.975 (21.722) * GARCH(β) 0.956 (139.108) * 0.828 (64.177) * 0 (0.00) α+β 0.994 0.998 0.976 Panel C : Dynamic Conditional Correlation A 0.107 (8.330) * B 0.766 (25.166) * Log-Likelihood 10452.24 Panel D : Diagnostic Tests on Residuals Q (20) 42.624 (0.002) 22039 (0.000) 22110 (0.000) (*) Indicates the significance of values at the 5% threshold. Q (20): Represents the autocorrelation test statistics applied to the residuals. Table 5. Estimation of the DCC-GARCH Model (1.1) (Platinum, CO2, and Temperature Anomaly) Platinum Co2 Anomalies Panel A : Mean Equation Constant -0.013 (-0.398) 0.004 (17.979) * 0.386 (56.658) * Panel B : Variance Equation Constant 0.013 (1.906) 0 (0.000) 0.001 (9.663) * ARCH(α) 0.036 (7.018) * 0.170 (12.162) * 0.975 (21.661) * GARCH(β) 0.959 (247.527) * 0.828 (64.177) * 0 (0.000) α+β 0.996 0.998 0.976 Panel C : Dynamic Conditional Correlation A 0.096 (7.616) * B 0.787 (24.323) * Log-Likelihood 10431.74 Panel D : Residual Diagnostic Tests Q (20) 75.268 (0.000) 22039 (0.000) 22110 (0.000) (*) Indicates the significance of values at the 5% threshold. Q (20): Represents the autocorrelation test statistics applied to the residuals. The analysis of the returns of gold, silver, and platinum through the DCC-GARCH (1,1) model highlights the predominant influence of environmental factors, such as CO2 emissions and temperature anomalies, on the dynamics of these precious metals. The results, presented in Tables 3, 4, and 5, show that, although the constants associated with the returns of gold (0.010), silver (-0.017), and platinum (-0.013) are not significant, the constants for CO2 (0.000 for gold and silver, 0.004 for platinum) and temperature anomalies (0.386 for all three metals) are highly significant, indicating their crucial role in the evolution of the mean dynamics. This aligns with the work of Karmakar et al. (2023), who also emphasize the importance of environmental factors in predicting the returns of precious metals. In terms of variance, the ARCH (α) and GARCH (β) coefficients reveal a high persistence of volatility for the three precious metals studied. For gold, the coefficients are α = 0.051 and β = 0.937; for silver, α = 0.038 and β = 0.956; and for platinum, α = 0.036 and β = 0.959. These results indicate that past shocks have a moderate but persistent effect on current volatility, which is crucial for long-term forecasting. For CO2, the coefficients α = 0.170 and β = 0.828 also show volatility influenced by past shocks, while temperature anomalies present a very high α (0.975) with no long-term persistence (β = 0). The sum of the coefficients α + β, close to 1 for all variables (0.988 for gold, 0.994 for silver, 0.995 for platinum, 0.998 for CO2, and 0.976 for temperature anomalies), confirms a high persistence of volatility, essential for long-term forecasts. In terms of dynamic conditional correlation, the returns of these metals and the environmental factors show correlations influenced by past shocks and persistent over time, with coefficients A = 0.117 and B = 0.744 for gold, A = 0.107 and B = 0.766 for silver, and A = 0.121 and B = 0.781 for platinum. This complex dynamic justifies the use of the DCC-GARCH model to capture these interdependencies, which are crucial for investors and risk managers. Finally, diagnostic tests on the residuals reveal significant autocorrelation, particularly for CO2 and temperature anomalies. This analysis highlights the importance of environmental factors in predicting the returns of precious metals. The persistence of volatility and the significant dynamic correlations between the returns and environmental variables make these factors essential for forecasting models. For investors, integrating these elements into their investment strategies is crucial in a context of growing environmental concerns, thus providing valuable insights for risk management and the development of sustainable investment strategies. The analysis of the accumulated impulse response functions, presented in Figure 1, reveals complex dynamics between carbon emissions, temperature anomalies, and precious metal returns before, during, and after the COVID-19 crisis. For gold, a CO2 shock before the pandemic led to a decrease in returns, illustrating a marked sensitivity to environmental pressures. However, after the pandemic, a reversal of this trend occurred, with an increase in returns, suggesting a possible adaptation or market response to a new post-crisis reality. Temperature anomalies generally had a negative effect on gold, with the notable exception of January 2, 2019, when a positive impact was observed. This particular date could correspond to exceptional circumstances or specific events that disrupted the usual relationship between temperature and the gold market. Regarding platinum, CO2 shocks continued to have a negative impact on its returns throughout the studied periods, although volatility decreased after the crisis, suggesting some market stabilization. Responses to temperature anomalies followed a similar trajectory, with a marked negative return during the pandemic, followed by a positive recovery afterward, indicating potential resilience in the platinum market to climate disruptions. Silver, on the other hand, exhibited particular resilience after the pandemic, with returns increasing in response to CO2 shocks, in contrast to other periods where the impact was negative. During the crisis, a temperature shock led to an increase in returns, suggesting a dynamic and potentially speculative response from investors to climate and health uncertainties. This analysis highlights the importance of environmental factors in determining precious metal returns, while underscoring the variability of responses depending on the economic and health periods. The recovery observed in the returns of gold and silver after the COVID-19 crisis may suggest that markets have adapted to new environmental risks, while platinum appears less responsive, with persistent negative returns despite reduced volatility. These findings open up interesting prospects for investment strategies that take into account evolving climate factors and their influence on financial assets. In line with the work of Miao et al. (2022), which emphasizes that external trends influence asset prices, these results suggest that the variability in the observed responses could be attributed to dynamics specific to each metal. These conclusions encourage further exploration of the predictive capacity of precious metal returns by more finely integrating carbon emissions and temperature anomalies into risk assessment models. Table 6. Forecasts from the TVC-VAR Model Performance Measures Variables RMSE MAE Theil Gold 0.00880 0.0063 0.858 Silver 0.0178 0.0129 0.863 Platinum 0.0181 0.0122 0.889 Table 6 highlights the performance of the TVC-VAR model for forecasting the returns of gold, silver, and platinum based on three key indicators: RMSE (Root Mean Square Error), MAE (Mean Absolute Error), and the Theil U coefficient. First, gold stands out with an RMSE of 0.0088 and an MAE of 0.0063, suggesting strong model performance in forecasting its returns. The Theil U coefficient for gold is 0.858, indicating a good match between the forecasts and the actual values. This underscores the relative stability of gold, often regarded as a safe-haven asset during economic turbulence. This stability allows investors to have some confidence in the model's predictions for gold, making it a strategic choice for those seeking predictable returns in an uncertain environment. For silver, although the RMSE is slightly higher (0.0178) as well as the MAE (0.0129), the forecasts remain robust. The Theil U coefficient for silver shows that the model captures the dynamics of this metal well, although it is more sensitive to economic and industrial fluctuations. Due to its numerous industrial applications, silver often reacts to supply and demand shocks on a global scale. These results suggest that its returns can be anticipated with some degree of confidence, while keeping in mind that unexpected movements may occur. Investors might consider using these forecasts within a diversified strategy to mitigate the risks associated with silver. Regarding platinum, the RMSE (0.0181) and MAE (0.0122) indicate that the model captures its fluctuations well, though the forecasts reveal slightly higher variability. The Theil U coefficient of 0.889 reflects the model's ability to forecast platinum returns, even though this metal is influenced by specific factors such as environmental regulations and technological advances, particularly in the automotive industry. This means that investors interested in platinum may need to be more vigilant and adopt a more dynamic approach, considering external factors that could alter platinum's returns in the medium and long term. Thus, the results in Table 6 show that the TVC-VAR model provides relevant forecasts for precious metals, with varying performances depending on the nature of the metal. Gold, with its stability, is particularly well-anticipated by the model, making it a strategic choice for investors seeking reliable forecasts. Silver and platinum, while presenting some volatility, are also satisfactorily modeled, but require more active management to leverage market dynamics. Finally, these results highlight the importance of environmental variables in forecasting precious metal returns, calling for increased consideration of these factors in investment decisions. These findings contrast with the work of Urquhart (2017), which identifies platinum as the most predictable, and Cohen (2022), who emphasizes the predictability of silver. This divergence suggests that accounting for environmental factors in modeling could alter the predictability order of precious metals and enrich the understanding of market dynamics. In this perspective, several recommendations emerge for market players in the precious metals sector, each tailored to the specific needs of the different groups involved. For investors, it is crucial to diversify portfolios by including precious metals such as gold and silver, which offer relative stability and predictable returns in uncertain environments. However, the increased volatility of platinum requires a more nuanced approach, where investors must consider specific industrial and regulatory factors influencing this metal. For example, strict environmental regulations in the automotive industry could affect the demand for platinum, making its valuation more complex. The adoption of advanced econometric models, such as DCC-GARCH and TVC-VAR, can help investors better anticipate return fluctuations and manage associated risks. These models are particularly useful for capturing the effects of environmental shocks, such as CO2 emissions and temperature anomalies, which have a significant impact on precious metal markets. Furthermore, given the risks associated with negative asymmetry and leptokurtosis observed in returns, proactive risk management is essential. Investors can use derivative instruments, such as options or futures contracts, to protect themselves against extreme events and limit potential losses. For policymakers and regulators, it is essential to implement public policies aimed at mitigating the impacts of climate change on financial markets. For example, measures to reduce CO2 emissions, such as carbon taxes or subsidies for green technologies, can have direct effects on the prices of precious metals. Moreover, raising awareness among market participants about climate risks and their influence on metal prices is crucial to promoting sustainable investment strategies. Regulators could also encourage academic research on the interactions between environmental factors and financial markets, for example, by funding interdisciplinary studies combining finance, economics, and climatology. These efforts would contribute not only to a better understanding of market dynamics but also to the development of regulatory frameworks adapted to current environmental challenges. For companies in the precious metals sector, it is crucial to adopt more sustainable production strategies to meet the growing demands of consumers and regulators. For instance, investing in clean technologies to reduce CO2 emissions related to metal extraction and refining could enhance their competitiveness in the market. Additionally, companies need to anticipate the impacts of climate anomalies on their supply chains and adapt their operations accordingly. For example, extreme weather events could disrupt mining activities, requiring robust continuity plans. Collaborating with researchers to explore sustainable alternatives to traditional precious metals could also open up new business opportunities. Finally, for financial analysts and portfolio managers, it is essential to explicitly integrate environmental variables into their forecasting models. For example, including data on CO2 emissions and temperature anomalies in analyses could improve the accuracy of predictions and enable better anticipation of future trends. These professionals should also develop adaptive investment strategies that can quickly adjust to changes in climatic and regulatory conditions. For instance, overweighting assets less sensitive to environmental shocks or underweighting those exposed to strict regulations could optimize portfolio performance. In sum, these recommendations aim to better address the risks and opportunities associated with precious metals in the context of ecological transition and increasing environmental regulations. They emphasize the importance of a proactive and interdisciplinary approach to navigating the complex challenges posed by the interactions between finance and the environment. 5. Conclusion The primary goal of this study was to examine the influence of climate variations and CO2 emissions on precious metal returns, specifically gold, silver, and platinum, while offering a dynamic and updated perspective compared to previous research. The results enrich our understanding of the complex relationships between environmental concerns and financial markets. They reveal that environmental factors profoundly alter the traditional perception of the predictability of precious metals. This analysis emphasizes the importance of integrating systemic risks related to climate change into investment strategies, opening new pathways for portfolio management and market regulation. A complex dynamic emerges between environmental factors and the predictability of returns. The analysis of returns through the DCC-GARCH and TVC-VAR models highlights this relationship. Gold, with relatively stable volatility (α = 0.051, β = 0.937) and solid forecasts (RMSE = 0.0088, MAE = 0.0063), maintains its status as a reliable safe-haven asset. However, its growing sensitivity to CO2 emissions, as shown by the impulse response functions before and after the COVID-19 pandemic, indicates that environmental concerns may affect its traditional role as a safe-haven in the context of a climate crisis. In contrast, silver and platinum exhibit more volatile behaviors and require active management due to their high responsiveness to environmental and economic shocks. Platinum, in particular, shows increased persistence in its correlations with climate anomalies (A = 0.096, B = 0.787), while silver demonstrates post-pandemic resilience to CO2 shocks. These findings highlight that considering environmental factors is essential for anticipating market fluctuations in the context of ecological transition. The ARCH and GARCH coefficients reveal strong persistence of past shocks on current volatility for all three metals, confirming the importance of these factors for long-term forecasting. The sum of the coefficients α + β, close to 1 for all variables (0.988 for gold, 0.994 for silver, 0.995 for platinum, 0.998 for CO2, and 0.976 for temperature anomalies), illustrates strong inertia in volatility, crucial for long-term forecasts. The dynamic conditional correlations captured by the DCC-GARCH model reveal crucial interdependencies between precious metal returns and environmental variables. These interdependencies fully justify the use of advanced models to better understand the risks and opportunities associated with precious metals. Additionally, the accumulated impulse response functions show varied adjustments of precious metals to environmental shocks before, during, and after the COVID-19 pandemic, reflecting a progressive adaptation of markets to new economic and climate realities. The TVC-VAR model forecasts confirm these differentiated dynamics. Gold stands out for its relative stability and increased predictability, making it a strategic choice for investors seeking predictable returns. However, its sensitivity to CO2 emissions underscores the need to incorporate environmental concerns into investment strategies. Platinum, on the other hand, exhibits a stronger responsiveness to environmental shocks, requiring proactive management to mitigate risks associated with its dual role as an industrial and precious metal. Silver, although affected by similar factors, shows post-pandemic resilience and volatility linked to its industrial applications, which calls for a diversified approach to maximize returns. In conclusion, this study underscores the importance of integrating environmental concerns, such as CO2 emissions and climate anomalies, into the analysis of precious metal returns. It urges investors to adopt more dynamic and diversified approaches to better anticipate market fluctuations in the context of ecological transition. Regulators also play a key role in fostering policies that support the energy transition and the sustainable use of precious metals. Finally, these results open interesting avenues for future research on the impact of environmental policies and climate shocks on financial markets, as well as the integration of these factors into more robust economic models. In this context, several recommendations emerge for stakeholders in the precious metals market. For investors, it is recommended to diversify their portfolios with metals such as gold and silver while managing platinum's volatility using advanced models like DCC-GARCH and TVC-VAR. The use of derivatives is essential to mitigate risks associated with extreme events. Policymakers should promote policies that reduce CO2 emissions and raise awareness of the connection between climate and the metals market. Analysts and companies must integrate environmental factors into their models and strategies to anticipate regulatory and climate developments. These recommendations aim to better understand the risks and opportunities associated with precious metals in the context of ecological transition and increasing environmental regulations. Declarations Author Contribution Ilyes Abidi: Was responsible for the empirical analysis, including the selection of methodology and econometric modeling. Ilyes developed the analytical framework, conducted the statistical analysis, and interpreted the results.Maissa Mejri: Contributed to data collection, the development of the theoretical framework, and the literature review. Maissa played a key role in gathering data and ensuring a solid theoretical foundation for the study.Mariem Nsaibi: Provided supervision and guidance throughout the research process. Mariem contributed to the overall direction and coherence of the manuscript, providing valuable editorial revisions to enhance clarity.Kamel Touhami: Contributed supervision, guidance, and editorial revisions to improve the manuscript’s clarity and coherence. Kamel provided insightful feedback to strengthen the quality of the manuscript. Data Availability The data used in this study were sourced from the following publicly available databases:• https://www.macrotrends.net/• https://gml.noaa.gov/• https://climatechange.umaine.edu/climate-matters/climate-reanalyzer/#:~:text=Climate%20Reanalyzer%20is%20a%20platform,data%20easily%20accessible%20by%20anyone. References Abidi I, Touhami K (2024) Safe haven for crude oil: Bitcoin or precious metals? New insight from time varying coefficient-vector autoregressive model. Int J Energy Econ Policy 14(1):184–195. https://doi.org/10.32479/ijeep.15107 Addison T, Ghoshray A (2023) Discerning trends in international metal prices in the presence of nonstationary volatility. Resour Energy Econ 71:101334. https://doi.org/10.1016/j.reseneeco.2022.101334 Baur DG, Lucey BM (2010) Is gold a hedge or a safe haven? An analysis of stocks, bonds and gold. Financial Rev 45(2):217–229. https://doi.org/10.1111/j.1540-6288.2010.00244.x BlackRock (2021) Investing with purpose 2021 annual report. https://s24.q4cdn.com/856567660 Cohen G (2022) Algorithmic strategies for precious metals price forecasting. Mathematics 10(7):1134. https://doi.org/10.3390/math10071134 Dichtl H (2020) Forecasting excess returns of the gold market: Can we learn from stock market predictions? J Commod Markets 19:100106. https://doi.org/10.1016/j.jcomm.2019.100106 Dinh T, Goutte S, Nguyen DK, Walther T (2022) Economic drivers of volatility and correlation in precious metal markets. J Commod Markets 28:100242. https://doi.org/10.1016/j.jcomm.2021.100242 Engle R (2002) Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models. J Bus Economic Stat 20(3):339–350. https://doi.org/10.1198/073500102288618487 Huang J, Cao Y, Zhong P (2022) Searching for a safe haven to crude oil: Green bond or precious metals? Finance Res Lett 50:103303. https://doi.org/10.1016/j.frl.2022.103303 International Council on Mining and Metals (2019) Integrated mine closure. https://www.icmm.com/integrated-mine-closure Karmakar S, Gupta R, Cepni O, Rognone L (2023) Climate risks and predictability of the trading volume of gold: Evidence from an INGARCH model. Resour Policy 82:103438. https://doi.org/10.1016/j.resourpol.2023.103438 Mehrdoust F, Noorani M (2024) Forecasting precious metals price based on artificial neural network trained by Lévy flight optimization algorithm. https://doi.org/10.2139/ssrn.4722754 . SSRN Metals Focus (2021) Reports. https://www.metalsfocus.com/reports/ Miao M, Khaskheli A, Raza SA, Yousufi SQ (2022) Using internet search keyword data for predictability of precious metals prices: Evidence from non-parametric causality-in-quantiles approach. Resour Policy 75:102478. https://doi.org/10.1016/j.resourpol.2021.102478 Nguyen D, Walther T (2020) Modeling and forecasting commodity market volatility with long-term economic and financial variables. J Forecast 39(2):126–142. https://doi.org/10.1002/for.2616 Papenfuß P, Schischke A, Rathgeber A (2025) Factors of predictive power for metal commodities. North Am J Econ Finance 76:102309. https://doi.org/10.1016/j.najef.2024.102309 Pierdzioch C, Risse M (2020) Forecasting precious metal returns with multivariate random forests. Empirical Economics 58:1167–1184. https://doi.org/10.1007/s00181-018-1558-9 Raza SA, Masood A, Benkraiem R, Urom C (2023) Forecasting the volatility of precious metals prices with global economic policy uncertainty in pre and during the COVID-19 period: Novel evidence from the GARCH-MIDAS approach. Energy Econ 120:106591. https://doi.org/10.1016/j.eneco.2023.106591 Research and Markets (2022) Global electric vehicle market analysis and forecast to 2027. https://www.researchandmarkets.com/ Shahid MN, Jehanzeb M, Abbas A, Zubair A, Akbar MAH (2020) Predictability of precious metals and adaptive market hypothesis. Int J Emerg Markets 15(5):1011–1027. https://doi.org/10.1108/IJOEM-07-2018-0404 Swamy PAVB, Hall SG, Tavlas GS, Hondroyiannis G (2010) Estimation of parameters in the presence of model misspecification and measurement error. Stud Nonlinear Dyn Econom 14(1):1–33. https://doi.org/10.2202/1558-3708.1743 Urquhart A (2017) How predictable are precious metal returns? Eur J Finance 23(14):1390–1413. https://doi.org/10.1080/1351847X.2016.1204334 World Gold Council (2020) Gold demand trends full year and Q4 2020. https://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-and-q4-2020/15507 World Gold Council (2022) Gold demand trends: Full year 2022. https://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-2022 Yang J, Agyei SK, Bossman A, Gubareva M, Marfo-Yiadom E (2024) Energy, metals, market uncertainties, and ESG stocks: Analysing predictability and safe havens. North Am J Econ Finance 69:102030. https://doi.org/10.1016/j.najef.2023.102030 Zhang Y-J, Pan X (2021) Does the risk aversion of crude oil market investors have directional predictability for the precious metal and agricultural markets? China. https://doi.org/10.1108/CAER-08-2020-0275 . Agricultural Economic Review Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6099418","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":425555406,"identity":"ce0d9e44-6abf-4d2a-903f-08caa464479e","order_by":0,"name":"Ilyes Abidi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYDACCcYGiQQGBhkgk/ExWISZuYGhgggtPCC1xgwMBkCKsYHhDF4tYATWwiYN1sJAQAv/7ObGGw8YbHgMjrc/qy6o+BPN3w7UcnAPHkvuHGy2SGBI4zE4c8bs9owzBrkzDgO1HHiGx5obiW1AvxzmMbiRw3abt80gtwGohfnDAdw65CFa/vMY3H/+rBikZT7YFjxaDCBaDgBtYTBjBmnZQEiLIdgvBsk8kmdyjKV5zhjnbgRqOYBPi9zt9oc3f1TYyfEdP/7wM0+FXO6884cPPsCnBeo8BgYFZEUENYCBfANRykbBKBgFo2AkAgAgflkx5LjuqwAAAABJRU5ErkJggg==","orcid":"","institution":"Management Information Systems Department, Applied College, University of Ha’il, Hail City P.O. Box 2440, Saudi Arabia.","correspondingAuthor":true,"prefix":"","firstName":"Ilyes","middleName":"","lastName":"Abidi","suffix":""},{"id":425555408,"identity":"caf7a123-fa09-4bad-95e4-c2662d339624","order_by":1,"name":"Kamel Touhami","email":"","orcid":"","institution":"Faculty of Economics and Management of Nabeul, University of Carthage, Tunisia. University Campus Mrezga route Hammamet 8000, Nabeul, Tunisia.","correspondingAuthor":false,"prefix":"","firstName":"Kamel","middleName":"","lastName":"Touhami","suffix":""},{"id":425555410,"identity":"8957520f-cc80-4118-810f-46383c814614","order_by":2,"name":"Mariem Nsaibi","email":"","orcid":"","institution":"Management Information Systems Department, Applied College, University of Ha’il, Hail City P.O. Box 2440, Saudi Arabia.","correspondingAuthor":false,"prefix":"","firstName":"Mariem","middleName":"","lastName":"Nsaibi","suffix":""},{"id":425555411,"identity":"d0b2e2db-6587-452b-a7c1-b1ead9cc9d68","order_by":3,"name":"Maissa Mejri","email":"","orcid":"","institution":"Faculty of Economics and Management of Nabeul, University of Carthage, Tunisia. University Campus Mrezga route Hammamet 8000, Nabeul, Tunisia.","correspondingAuthor":false,"prefix":"","firstName":"Maissa","middleName":"","lastName":"Mejri","suffix":""}],"badges":[],"createdAt":"2025-02-24 19:08:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6099418/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6099418/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":78224405,"identity":"65f7daa2-2287-46ef-85a3-be2c63441b5a","added_by":"auto","created_at":"2025-03-11 06:48:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":368352,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAccumulated Response Function of Precious Metal Returns to Carbon Emissions and Temperature Anomalies Before, During, and After COVID-19\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6099418/v1/666ef492c3c0b4c2da03d91c.png"},{"id":78229265,"identity":"1511c79b-8b4f-4a5d-a588-4febf9f5e808","added_by":"auto","created_at":"2025-03-11 07:20:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1481010,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6099418/v1/95174f67-a6aa-4ec9-a288-0d47c11b2865.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Can Environmental Factors Predict Precious Metal Returns?","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn an ever-evolving global economic context marked by geopolitical crises, financial tensions, and a rapid transition to sustainable economic models, precious metals play a key role in investors' portfolios looking to secure their wealth. Historically seen as safe havens, these assets continue to be valued for their ability to preserve wealth in times of economic uncertainty, particularly amid financial market volatility, persistent inflation, and geopolitical disruptions. In 2022, gold reached a record global demand, approaching 4,700 tons, highlighting its essential role in risk management (World Gold Council, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, beyond traditional economic factors, environmental elements are increasingly influencing the dynamics of precious metal markets.\u003c/p\u003e \u003cp\u003eThe extraction of precious metals, while fundamental to global supply, presents significant environmental challenges. For instance, producing just one gram of gold can generate up to 20 tons of mining waste, resulting in considerable CO2 emissions and damage to local biodiversity (ICMM, 2019). In response to these concerns, stricter regulations have been imposed, increasing extraction costs for some gold mines by up to 30% (World Gold Council, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Consequently, there has been a surge in demand for investments in responsible mining companies, with more than \u003cspan\u003e$\u003c/span\u003e50\u0026nbsp;billion in assets invested in 2020 (BlackRock, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). These developments illustrate the growing pressure for more sustainable production, supported by initiatives like the \"Responsible Gold Mining Principles\" from the Council on Mining and Metals (ICMM).\u003c/p\u003e \u003cp\u003eAlongside this trend, the transition to a green economy is reshaping the demand for these resources. Precious metals, seen as sustainable materials, now play a central role in discussions about energy transition and the environmental impact of mining. Economic actors must rethink their understanding of market dynamics, incorporating factors such as climate risks, ESG (Environmental, Social, and Governance) policies, and technological innovations in renewable energy.\u003c/p\u003e \u003cp\u003ePrevious research on the predictability of precious metal returns has mainly focused on traditional economic factors such as interest rates, inflation, and geopolitical shocks. Urquhart (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) highlighted the complexity of short-term prediction of these assets, often regarded as safe havens, due to the unpredictable volatility of markets. In this line, Dichtl (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) demonstrated that incorporating long-term financial indicators improves the robustness of forecasts, though they remain limited by the high sensitivity of precious metals to financial market fluctuations. Taking a broader perspective, Papenfu\u0026szlig; et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) analyzed the evolution of predictive factors for precious metals and found that the influence of autoregressive price components has gradually diminished, with financial market indices becoming more significant, particularly after the 2008 crisis.\u003c/p\u003e \u003cp\u003eIn light of the limitations of traditional approaches, Shahid et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) applied the AMH (Adaptive Market Hypothesis) theory to precious metals and demonstrated the temporal variability in their predictability, with investors continuously adjusting their strategies based on new economic and environmental information. This dynamic is particularly pronounced during crises, making classical models less effective. In this context, Zhang and Pan (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) highlighted the impact of economic and geopolitical crises on the volatility of precious metal prices, especially in relation to oil market fluctuations. Similarly, Huang et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) confirmed gold's safe-haven role, while tempering this function for silver and platinum, which are more exposed to industrial demand fluctuations.\u003c/p\u003e \u003cp\u003eThe importance of economic uncertainties was also highlighted by Raza et al. (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) through the GARCH-MIDAS approach, which integrates global macroeconomic variables and improves forecast accuracy. At the same time, Yang et al. (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) showed that while gold and oil are influential assets, they do not serve as safe havens for ESG assets but contribute to portfolio diversification. In a complementary approach, Karmakar et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) used an INGARCH model to analyze the impact of climate risks on trading volumes of gold futures, revealing that physical risks significantly influence these volumes in the short term.\u003c/p\u003e \u003cp\u003eThese advances illustrate that the predictability of precious metal returns is constantly evolving, influenced by structural changes and new investment paradigms. The integration of environmental risks and adaptive approaches is therefore crucial for refining predictive models. However, these models still face major challenges, particularly with climate shocks that directly disrupt precious metal production. For example, in 2021, extreme weather events reduced gold production by 15% in certain regions (Metals Focus, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), highlighting the industry's vulnerability to environmental disruptions. These factors emphasize the need to integrate climate risks more thoroughly into analyses to better anticipate long-term impacts on the precious metal market.\u003c/p\u003e \u003cp\u003eAt the same time, the demand for precious metals used in green technologies, such as electric vehicles and solar panels, is growing rapidly. According to Research and Markets (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), the electric vehicle market, reliant on palladium, platinum, and silver, is expected to reach a value of \u003cspan\u003e$\u003c/span\u003e800\u0026nbsp;billion by 2027. These transformations in the demand and supply of precious metals call for a revision of traditional forecasting models, which struggle to integrate these new dynamics.\u003c/p\u003e \u003cp\u003eThis study aims to examine the impact of environmental factors on the predictability of precious metal returns, specifically focusing on gold, silver, and platinum, as well as environmental variables such as CO2 emissions and temperature anomalies. The main objective is to analyze how these environmental variables influence price volatility and evaluate how markets respond to risks associated with the energy transition, while providing methodological tools to help investors better integrate these environmental factors into their investment decisions.\u003c/p\u003e \u003cp\u003eThe methodological framework relies on two advanced econometric models: the DCC-GARCH (Dynamic Conditional Correlation Generalized Autoregressive Conditional Heteroskedasticity) model and the TCV-VAR (Time-Varying Coefficients Vector Autoregression) model. The DCC-GARCH model is used to analyze the dynamic conditional correlations between precious metal returns and environmental factors, offering a detailed understanding of market adjustments to ecological changes and energy transition policies. The TCV-VAR model, on the other hand, allows for the examination of the dynamic interdependencies between these variables and captures the time-varying nature of relationships between precious metal returns and environmental variables by modeling nonlinear and time-changing effects. This ability to adapt coefficients over time is crucial for capturing the short- and long-term impacts of environmental shocks.\u003c/p\u003e \u003cp\u003eBoth models are estimated using daily data covering the period from January 5, 2017, to October 2, 2023, allowing for an in-depth analysis of temporal dependencies and conditional heteroskedasticity, essential for obtaining reliable results in this study context.\u003c/p\u003e \u003cp\u003eWhile classical econometric models like GARCH-MIDAS, applied by Nguyen and Walther (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Dinh et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), have evaluated the long-term impact of macroeconomic factors on the volatility of precious metals, these approaches have notable limitations. The GARCH-MIDAS mainly focuses on integrating long-term macroeconomic variables without considering structural changes in coefficients over time. Furthermore, AI-based approaches, such as those proposed by Pierdzioch and Risse (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Cohen (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), have indeed improved the prediction of returns but suffer from a lack of economic interpretability, making it difficult to understand the underlying mechanisms between financial markets and environmental variables. The use of alternative data, such as internet search trends explored by Miao et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), remains limited as it cannot effectively assess interactions with specific environmental indicators.\u003c/p\u003e \u003cp\u003eIn response to these gaps, our research stands out by explicitly integrating environmental factors into the predictability of precious metal returns. The TCV-VAR model allows for analyzing the dynamic evolution of relationships between precious metal returns and environmental variables, capturing nonlinear and time-varying effects. Unlike classical models, it incorporates the adaptability of coefficients over time, providing a better understanding of the short- and long-term impacts of environmental shocks. Additionally, the DCC-GARCH model is used to evaluate the dynamic conditional correlations between precious metal returns and environmental indicators, offering a nuanced analysis of market adjustments to ecological changes and energy transition policies.\u003c/p\u003e \u003cp\u003eBy combining these two approaches, our study fills a major methodological gap in the existing literature. It provides a more comprehensive and nuanced perspective on the influence of environmental factors on the dynamics of precious metals while meeting the needs of investors, regulators, and policymakers in the context of the transition to a greener economy.\u003c/p\u003e \u003cp\u003eThis paper is structured as follows. Section \u003cspan refid=\"Sec2\" class=\"InternalRef\"\u003e2\u003c/span\u003e offers a review of relevant literature. Section \u003cspan refid=\"Sec3\" class=\"InternalRef\"\u003e3\u003c/span\u003e details the methodological framework and data analysis. The main results are discussed in Section \u003cspan refid=\"Sec6\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Finally, Section \u003cspan refid=\"Sec7\" class=\"InternalRef\"\u003e5\u003c/span\u003e presents the conclusions and implications of this study.\u003c/p\u003e"},{"header":"2. Literature Review","content":"\u003cp\u003ePrecious metals, due to their historical role as a store of value, continue to attract the attention of investors and researchers. The predictability of their returns has been the subject of numerous studies, particularly during periods of high financial market volatility. While traditional economic factors, such as inflation, interest rates, and financial crises, have long been considered the primary determinants of these returns, more recent research has highlighted the growing importance of environmental, social, and governance (ESG) variables. At the same time, the rise of advanced quantitative methods and artificial intelligence technologies has significantly improved the forecasting capabilities of existing models. These developments offer a better understanding of the mechanisms influencing precious metal prices and open new perspectives for analyzing their behavior in financial markets.\u003c/p\u003e \u003cp\u003eHistorically, the returns of precious metals have been heavily influenced by macroeconomic variables and global economic cycles. Urquhart (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) analyzes the predictability of precious metal returns through classical econometric models (Variance ratio test, Brock, Dechert, and Scheinkman test, Hurst exponent) and emphasizes that, although these assets are often considered safe havens, their short-term predictability remains complex. Price fluctuations can be exacerbated by unforeseen events, such as financial crises, making predictions uncertain. In this regard, Dichtl (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) explores the possibility of using methodologies from stock markets to forecast excess returns of gold. He shows that models incorporating long-term financial indicators provide more robust forecasts, but these remain limited by gold's high sensitivity to financial market dynamics. Thus, these studies highlight that the stability of financial markets plays a determining role in the predictability of precious metals, but that this predictability is conditioned by the ability of models to integrate often unpredictable external factors.\u003c/p\u003e \u003cp\u003eIn this context, Papenfu\u0026szlig; et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) examine the evolution of predictive factors for metal prices in a context marked by phases of financialization and de-financialization of commodity markets. Through the analysis of 24 metals over the period 1995\u0026ndash;2019, they demonstrate that the autoregressive components of prices are the primary determinants, although their influence gradually decreases. They also highlight the importance of interest rates before the 2008 financial crisis, while financial market indices play a more dominant role afterward. In terms of predictive performance, their models significantly outperform traditional benchmarks, such as the random-walk model, in 12 out of the 24 cases studied, particularly for minor metals. These results confirm that the relationships between macroeconomic variables and metal prices are dynamic and evolve over time, justifying the use of flexible forecasting models adapted to the structural changes of the market.\u003c/p\u003e \u003cp\u003eIn the face of these challenges, some researchers have adopted an alternative approach based on the Adaptive Market Hypothesis (AMH), which considers that investors adjust their strategies according to new economic and environmental information. Shahid et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) apply this approach to precious metals and show that their returns evolve according to an adaptive process, reflecting investors' constant adjustments in response to economic fluctuations and market shocks. This theoretical framework helps explain why return predictability is not a fixed characteristic but varies over time depending on macroeconomic conditions and market expectations. This adaptability is especially pronounced in times of crisis, where uncertainty amplifies investors' adjustments, making traditional models less effective.\u003c/p\u003e \u003cp\u003eAnother key aspect of this predictability is the impact of economic and geopolitical crises on precious metals. Zhang and Pan (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) show that the price volatility of gold and platinum is significantly influenced by fluctuations in the oil market, particularly during periods of global economic tension. Financial crises, such as the 2008 crisis or the COVID-19 pandemic, have strengthened the appeal of precious metals as safe-haven assets, leading to an increase in their returns. Huang et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) confirm this trend by analyzing the ability of precious metals to serve as a hedge against geopolitical and economic risks. However, they note that this safe-haven function varies across metals, with gold being historically more resilient than silver or platinum, which remain more susceptible to industrial fluctuations. In this perspective, understanding the dynamics of precious metal volatility during periods of economic uncertainty becomes essential. Modeling this volatility is thus a central issue, allowing for price movements to be anticipated in response to macroeconomic shocks.\u003c/p\u003e \u003cp\u003eRaza et al. (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) apply the GARCH-MIDAS approach to examine the impact of global economic uncertainties on precious metal volatility, particularly during the COVID-19 pandemic. Their findings show that incorporating global macroeconomic variables improves forecasting accuracy, highlighting the sensitivity of precious metals to economic uncertainty shocks. While precious metals have long been favored for their role as protection against economic crises, more recent financial instruments, such as green bonds, have begun to emerge as viable alternatives. Huang et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) analyze the effectiveness of green bonds as a hedge against economic shocks and conclude that they offer more sustainable stability in the face of environmental and social risks. Unlike precious metals, which strongly react to economic cycles and financial crises, green bonds allow for more effective diversification in long-term investment contexts. These findings open new perspectives for investors seeking to integrate ESG criteria into their strategies, while questioning the role of precious metals in a changing financial environment.\u003c/p\u003e \u003cp\u003eIn this regard, Yang et al. (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) examine the predictability of ESG stock returns, taking into account the role of traditional assets like gold and oil, as well as uncertainties related to the market, cryptocurrencies, and geopolitical risks. By applying non-parametric techniques such as quantile causality and quantile-on-quantile regression, their results show that gold, oil, market-implied volatility (VIX and OVX), and geopolitical risk are significant predictors of ESG stock returns. However, neither gold nor oil play the role of a safe haven for these assets but rather serve to diversify portfolios. On the other hand, ESG stocks appear to be an effective hedge against geopolitical shocks and uncertainties related to cryptocurrencies during bearish market periods. These findings underscore the importance of dynamic portfolio management for sustainable investments and strengthen the literature on the resilience of ESG stocks in the face of financial and environmental crises, while guiding investors on the strategic implications of these instruments in asset allocation and risk management.\u003c/p\u003e \u003cp\u003eMoreover, de Karmakar et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) explore the impact of both physical and transition climate risks on the trading volume of gold futures contracts. Due to the nature of the data, which is count-type, the authors use an INGARCH (log-linear Poisson GARCH-type) model to predict these volumes based on covariates related to climate risks. Their analysis reveals that physical risks have significant predictive power for gold trading volumes at 5-day and 22-day horizons. Additionally, a positive relationship between physical risks and gold trading volumes is observed, indicating that gold serves as a hedge against short-term physical risks (1 week and 1 month). Similar results are found for platinum and palladium, but not for silver. The authors emphasize the importance of gold as a safe-haven asset in the face of climate risks, especially physical risks, and demonstrate that this analysis provides a first direct approach to predicting the trading volumes of precious metals using count-based data models.\u003c/p\u003e \u003cp\u003eAt the same time, technological advancements have significantly transformed the methods for forecasting precious metal returns. Pierdzioch and Risse (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) explore the use of random forests, a supervised learning technique, to predict the returns of precious metals. Their results show that these models capture complex relationships between various economic variables and the prices of precious metals, offering an alternative to traditional econometric models. In a similar approach, Cohen (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) introduces advanced algorithmic strategies, based on artificial intelligence, to identify hidden patterns in the time series of precious metal prices. These emerging approaches pave the way for a significant improvement in forecasting capabilities, especially by incorporating external factors such as environmental and social data. Along the same lines, Mehrdoust and Noorani (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) introduce a forecasting model based on neural networks optimized by a L\u0026eacute;vy flight algorithm, enabling better handling of the complex price movements of precious metals. The results indicate that the neural network optimized by the L\u0026eacute;vy flight algorithm outperforms other prediction models in terms of accuracy. This model offers a unique approach to predicting precious metal prices and can be applied to different time series.\u003c/p\u003e \u003cp\u003eFurthermore, another promising approach involves leveraging online search trends to anticipate fluctuations in precious metal prices. Miao et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) study the impact of online search trends and show that these non-financial data can provide early signals about market developments. Using a non-parametric causality approach, they demonstrate that investor interest in specific keywords may be correlated with future price changes of precious metals. This new source of data, combined with machine learning models, could enhance investors' ability to anticipate market fluctuations.\u003c/p\u003e \u003cp\u003eFinally, the non-stationary volatility of precious metals remains a challenge for long-term forecasting. Addison and Ghoshray (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) explore this issue by applying models that incorporate changes in the variance of returns. Their results show that, while precious metals are often considered stable in the long term, they are subject to unpredictable variations, requiring adaptive models capable of absorbing these structural changes.\u003c/p\u003e \u003cp\u003eAlthough existing studies have provided significant insights into the predictability of precious metal returns, several limitations remain. On the one hand, many studies focus primarily on traditional macroeconomic factors such as inflation, interest rates, and financial market volatility, without giving sufficient attention to environmental factors. For example, the GARCH-MIDAS and DCC-GARCH models applied by Nguyen and Walther (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Dinh et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) have assessed the impact of long-term economic variables on precious metal volatility, but these models do not explicitly account for environmental factors. The GARCH-MIDAS, for instance, mainly focuses on integrating long-term macroeconomic factors but does not consider structural changes in coefficients over time. The same is true for the DCC-GARCH, which examines the dynamics of correlations between financial assets but does not adequately capture the dynamic impacts of environmental shocks or ecological policies on precious metal returns.\u003c/p\u003e \u003cp\u003eOn the other hand, the growing use of artificial intelligence and machine learning methods, as proposed by Pierdzioch and Risse (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) or Cohen (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), offers a notable improvement in forecasting returns. However, these approaches often suffer from a lack of interpretability and economic contextualization, limiting their applicability in understanding the underlying relationships between financial markets and environmental variables. Additionally, the study by Miao et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) on the use of online search trends to anticipate precious metal price variations highlights the importance of alternative data but does not assess their interaction with specific environmental indicators.\u003c/p\u003e \u003cp\u003eIn response to these limitations, our research stands out by explicitly integrating the role of environmental factors in the predictability of precious metal returns, using a combined approach of the Time-Varying Coefficients Vector Autoregression (TCV-VAR) model and the DCC-GARCH. The TCV-VAR model will allow us to analyze the dynamic evolution of relationships between precious metal returns and environmental variables, capturing non-linear and time-varying effects. Unlike classical models such as GARCH-MIDAS or DCC-GARCH, the TCV-VAR model allows for the modeling of the adaptability of variable coefficients over time, capturing both short-term and long-term effects of environmental shocks. Furthermore, the application of DCC-GARCH will enable us to assess the dynamics of conditional correlations between these returns and environmental indicators, providing a more nuanced understanding of how markets adjust to ecological changes and energy transition policies.\u003c/p\u003e \u003cp\u003eBy combining these two approaches, our study makes a significant contribution by bridging a methodological and conceptual gap in the existing literature. It enhances our understanding of how environmental factors influence the dynamics of precious metal returns and provides new tools for predicting price fluctuations in a rapidly evolving economic and ecological context.\u003c/p\u003e"},{"header":"3. Data and Methodology","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Data\u003c/h2\u003e \u003cp\u003eThis study aims to examine the impact of environmental factors on the predictability of precious metal returns, focusing specifically on gold, silver, and platinum, as well as environmental variables such as CO₂ emissions and temperature anomalies. The primary objective is to analyze how these environmental variables influence the volatility of precious metal prices and assess how markets respond to the risks associated with the energy transition. Additionally, the study provides methodological tools to assist investors in better integrating these environmental factors into their investment decisions.\u003c/p\u003e \u003cp\u003eTo address these questions, the empirical analysis relies on daily return data for three precious metals: gold, silver, and platinum, alongside daily data on carbon emissions and temperature anomalies. The return data for the precious metals were sourced from the Macrotrends website, a reputable provider of historical financial information. Regarding the environmental variables, carbon emissions were extracted from the Global Monitoring Laboratory, while temperature anomalies were retrieved from the Climate Reanalyzer platform. The data covers the period from January 5, 2017, to October 2, 2023.\u003c/p\u003e \u003cp\u003eThe returns for the precious metals are calculated from daily closing prices using the standard logarithmic return formula:\u003c/p\u003e \u003cp\u003eR\u003csub\u003et\u003c/sub\u003e=ln (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\frac{\\text{P}\\text{t}}{\\text{P}\\text{t}-1}\\:\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eWhere P\u003csub\u003et\u003c/sub\u003e and P\u003csub\u003et\u0026minus;1\u003c/sub\u003e represent the closing prices on consecutive days.\u003c/p\u003e \u003cp\u003eCarbon emissions and temperature anomalies are also expressed in terms of daily values, using them directly as provided by the aforementioned sources.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Methodology\u003c/h2\u003e \u003cp\u003eTo explore the impact of environmental factors on the predictability of precious metal returns, two econometric models are applied in this study: the DCC-GARCH (Dynamic Conditional Correlation Generalized Autoregressive Conditional Heteroskedasticity) model and the TCV-VAR (Time-Varying Coefficients Vector Autoregression) model.\u003c/p\u003e \u003cp\u003eThe DCC-GARCH model, introduced by Engle (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), is used to analyze the dynamics of the conditional correlation between the returns of precious metals and environmental variables (carbon emissions and temperature anomalies). This model captures the effects of conditional volatility and changes in correlation relationships over time, which is particularly relevant for financial time series. The model is estimated in two steps: first, univariate GARCH models are fitted to the return series, and then conditional correlations are estimated, accounting for volatility. The formula for the DCC-GARCH model can be described as follows:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eFirst step (univariate GARCH models):\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eY\u003csub\u003et\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;\u0026micro;\u003csub\u003et\u003c/sub\u003e + ϵ\u003csub\u003et​\u003c/sub\u003e\u003c/p\u003e \u003cp\u003eWhere y\u003csub\u003et\u003c/sub\u003e​ is the return, \u0026micro;\u003csub\u003et\u003c/sub\u003e is the conditional mean, and ϵ\u003csub\u003et\u003c/sub\u003e is the conditional error.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eSecond step (estimation of conditional correlations):\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eΣ\u003csub\u003et\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;Q + Aϵ\u003csub\u003et\u0026minus;1\u003c/sub\u003e ϵ\u0026prime;\u003csub\u003et\u0026minus;1\u003c/sub\u003e A\u0026prime; + BΣ\u003csub\u003et\u0026minus;1\u003c/sub\u003eB\u0026prime;\u003c/p\u003e \u003cp\u003eWhere Σ\u003csub\u003et\u003c/sub\u003e​ is the conditional variance-covariance matrix, Q is the constant matrix, and A and B are matrices that determine the volatility dynamics.\u003c/p\u003e \u003cp\u003eThe TCV-VAR model, formalized by Swamy et al. (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), is used to examine the dynamic interdependence between precious metal returns and environmental factors, while allowing the model's coefficients to vary over time. This enables the analysis of how the impact of environmental variables on precious metal returns evolves over different periods. The model is specified as follows:\u003c/p\u003e \u003cp\u003ey\u003csub\u003et\u003c/sub\u003e = C\u003csub\u003et\u003c/sub\u003e + \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sum\\:_{i=1}^{p}\\text{A}\\)\u003c/span\u003e\u003c/span\u003e\u003csub\u003ei\u003c/sub\u003ey\u003csub\u003et\u0026minus;i\u003c/sub\u003e + ϵ\u003csub\u003et\u003c/sub\u003e\u003c/p\u003e \u003cp\u003eWhere y\u003csub\u003et\u003c/sub\u003e ​is the vector of precious metal returns and environmental variables, C\u003csub\u003et\u003c/sub\u003e is the time-varying constant vector, A\u003csub\u003ei\u003c/sub\u003e represents the matrices of time-varying coefficients, and ϵ\u003csub\u003et\u003c/sub\u003e is the error vector.\u003c/p\u003e \u003cp\u003eBoth models are estimated using daily data and account for temporal dependencies and conditional heteroscedasticity effects, which is essential for obtaining reliable results in the context of a financial analysis of precious metals influenced by environmental factors.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Empirical Results and Discussion","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Descriptive Statistics of Returns\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"588\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGold\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSilver\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePlatinum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCO2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnomalies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000029\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e0.334\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMaximum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e6.789\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e8.896\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e11.176\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000743\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e0.753\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMinimum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e-5.400\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-12.345\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-12.315\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.000956\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e-0.832\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStandard Deviation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1.799\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1.762\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000312\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e0.224\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSkewness\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e-0.169\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.495\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.277\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.721\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e-1.717\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eKurtosis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e5.399\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e6.779\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e4.536\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.801\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e4.224\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eJarque-Bera\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e2116.160\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e3392.775\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1510.625\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e197.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e2141.759\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eProbability\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eObservations\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e1729\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1729\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1729\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1729\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e1729\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eQ\u0026sup2; (10)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e160.834\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e284.542\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e451.494\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e5669.857\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e7769.997\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 152px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eARCH (10)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e116.05\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e155.06\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e249.4\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1693.4\u003c/p\u003e\n \u003cp\u003e(0.000) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e1486.6\u003c/p\u003e(0.000) *\u003cbr\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e(*): indicates significance at the 5% level; Q\u0026sup2;(10) : are the statistics from the Ljung-Box test with 10 lags applied to squared returns; ARCH (10) : is the heteroscedasticity test by Engle (1982).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e presents the descriptive statistics of the returns for gold, silver, platinum, and environmental variables such as CO₂ emissions and temperature anomalies. The average returns reveal a slightly positive trend for gold (0.025) and silver (0.013), indicating a moderate appreciation of these precious metals over the studied period, while platinum shows a negative average (-0.003), suggesting a slight depreciation. These results are of great importance for investors wishing to diversify their portfolios with safe-haven assets. They align with numerous previous studies that indicate gold and silver often generate positive returns during periods of economic uncertainty. For example, the study by Baur and Lucey (2010) concludes that gold is a safe-haven asset, particularly during crises. However, our research stands out by including environmental factors, which seem to exert a more pronounced influence on the dynamics of precious metal returns, especially for platinum, which is more sensitive to industrial factors.\u003c/p\u003e\n\u003cp\u003eAs for volatility, the high standard deviations for silver (1.799) and platinum (1.762) compared to gold (0.868) highlight greater variability in returns, implying increased risk but also a higher potential for return, thereby attracting investors willing to tolerate higher volatility for potentially larger gains. The environmental variables, with much lower standard deviations (CO2: 0.000312; temperature anomalies: 0.224), show relative stability, reflecting the less volatile nature of these factors compared to precious metal returns.\u003c/p\u003e\n\u003cp\u003eThe negative skewness coefficients observed for gold (-0.169), silver (-0.495), platinum (-0.277), and environmental variables (CO2: -0.721; temperature anomalies: -1.717) indicate a left-skewed distribution of returns, meaning that significant losses are more frequent than substantial gains. This represents a significant downside risk for investors. This dynamic aligns with the work of Zhang and Pan (2021), who highlight that risk aversion influences return predictability, especially in the face of significant loss risks. Indeed, faced with negative skewness, investors are more likely to react negatively to the prospect of substantial losses, which can affect their perception of future returns and investment strategies. The interaction between this return skewness and risk aversion becomes even more critical in the context of precious metals, where environmental factors, such as CO₂ emissions and temperature anomalies, introduce additional uncertainties that may increase volatility and make returns even more unpredictable.\u003c/p\u003e\n\u003cp\u003eMoreover, the high kurtosis observed for gold (5.399) and silver (6.779) reveals a leptokurtic distribution, characterized by thick tails and an increased probability of extreme events. This feature emphasizes the importance for investors to protect themselves against extreme risks, using appropriate financial instruments or further diversifying their portfolios. This finding echoes that of Dinh et al. (2022), who show that stock market volatility can also induce extreme events in metal markets. Thus, the presence of high kurtosis in precious metal returns reinforces the idea that investors must not only account for daily fluctuations but also the risks linked to more severe and less frequent environmental shocks.\u003c/p\u003e\n\u003cp\u003eThe results of the Jarque-Bera normality tests, all significant (p-value = 0.000), confirm that the returns for precious metals and environmental variables do not follow a normal distribution. This deviation from normality suggests that traditional approaches based on the normality assumption may be inadequate for modeling and forecasting returns. In line with Dichtl (2020), who emphasizes that forecasting techniques must account for these deviations to be effective, it becomes essential to adopt advanced models capable of better capturing the underlying dynamics of returns. Therefore, the use of models such as DCC-GARCH and TVC-VAR is particularly relevant to capture the complexity of interactions between precious metal returns and environmental factors.\u003c/p\u003e\n\u003cp\u003eThe Q\u0026sup2; (10) and ARCH (10) statistics, both significant (p-value = 0.000), reveal the presence of autocorrelations and heteroscedasticity in the squared returns. This suggests the existence of volatility clusters, where periods of high volatility are followed by similar periods, a phenomenon typically observed in financial markets. For investors, this implies that periods of increased risk can be anticipated and managed more effectively using appropriate econometric models, thereby improving risk management and strategic decision-making.\u003c/p\u003e\n\u003cp\u003eBy integrating these analyses, it becomes clear that the returns of precious metals are influenced not only by traditional economic factors but also by environmental factors. Fluctuations in CO₂ emissions and temperature anomalies can affect the production and demand for precious metals, thus impacting their prices and returns. This interconnection highlights the importance for investors and policymakers to adopt a holistic approach in evaluating risks and opportunities in precious metal markets. In particular, considering environmental risks in investment strategies can lead to better portfolio diversification and reduced systematic risks associated with climate change and environmental regulations. These results are in line with the work of Huang et al. (2022), who emphasize the importance of external factors in the predictability of precious metal returns. However, our study deepens this perspective by more explicitly integrating environmental variables, thus highlighting their determining role in the evolution of precious metal markets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Unit Root Tests for Returns\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"577\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 255px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eADF Stationarity Test\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 246px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePP Stationarity Test\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"17\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo Constant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eWith Constant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eWith Constant and Trend\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo Constant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eWith Constant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eWith Constant and Trend\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"34\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd height=\"34\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"17\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGold\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-40.756\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-40.778\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-40.776\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-40.841\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e-40.908\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\n \u003cp\u003e-40.912\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"18\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSilver\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-42.409\u003c/p\u003e\n \u003cp\u003e(0.001) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-42.399\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-42.488\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-42.412\u003c/p\u003e\n \u003cp\u003e(0.001) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e-42.402\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\n \u003cp\u003e-42.391\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"18\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePlatinum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-41.344\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-41.332\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-41.320\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-41.864\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e-41.851\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\n \u003cp\u003e-41.837\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"18\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCo2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-2.922\u003c/p\u003e\n \u003cp\u003e(0.003) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-2.933\u003c/p\u003e\n \u003cp\u003e(0.041) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-2.932\u003c/p\u003e\n \u003cp\u003e(0.152)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-8.079\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e-8.132\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\n \u003cp\u003e-8.142\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"18\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnomalies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-1.558\u003c/p\u003e\n \u003cp\u003e(0.112)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-4.416\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-4.407\u003c/p\u003e\n \u003cp\u003e(0.002) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e-2.130\u003c/p\u003e\n \u003cp\u003e(0.031) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e-5.256\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\n \u003cp\u003e-5.243\u003c/p\u003e\n \u003cp\u003e(0.000) ***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"27\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 78px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"5\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;Note: The significance levels of 1%, 5%, and 10% are shown by ***, **, and *, respectively.\u003c/p\u003e\n\u003cp\u003eTable 2 presents the results of the unit root tests, specifically the Augmented Dickey-Fuller (ADF) test and the Phillips-Perron (PP) test, applied to the returns of gold, silver, platinum, as well as the two environmental variables: CO2 emissions and temperature anomalies.\u003c/p\u003e\n\u003cp\u003eThe results of the ADF and PP tests for the returns of the three precious metals show negative and statistically significant test statistics at the 1% level (p \u0026lt; 0.01), indicating that these series are stationary, regardless of whether a constant or trend is included. For example, for gold, the ADF values range from -40.756 to -40.776, and the PP values range from -40.841 to -40.912, all significant at the 1% level. This means that the returns of gold, silver, and platinum do not contain a unit root, implying that shocks affecting these series are temporary and do not persist indefinitely. This is a crucial factor for investors looking to model the volatility and future returns of these assets. From a financial perspective, this suggests that the returns of these precious metals revert quickly to their mean after a shock, providing some predictability in their behavior.\u003c/p\u003e\n\u003cp\u003eRegarding the environmental variables, the results are more mixed. For CO2 emissions, although the PP tests indicate strong stationarity with negative values (ranging from -8.079 to -8.142, all significant at the 1% level), the ADF test results are more nuanced. Without trend or constant, CO2 shows an ADF statistic of -2.922, significant at the 1% level, suggesting stationarity. However, when a constant and trend are included, the significance slightly decreases (p = 0.152), which could indicate the series\u0026apos; sensitivity to these components, potentially reflecting underlying trends not captured by the simple model. This highlights the importance of using a dynamic model to avoid incorrect conclusions.\u003c/p\u003e\n\u003cp\u003eRegarding temperature anomalies, establishing stationarity is more complex. The ADF results without trend or constant (-1.558) are not significant, suggesting that the series could be non-stationary. However, with the addition of a constant and trend, the ADF values (-4.416 and -4.407) become significant at the 1% level, indicating conditional stationarity. The PP test results reinforce this conclusion, with statistics ranging from -2.130 to -5.256, all significant except without constant or trend. This suggests that temperature anomalies may have a trend component or other complex dynamics that require more sophisticated modeling to fully capture their behavior. From a financial perspective, this could suggest that returns linked to temperature anomalies are influenced by long-term trends or exogenous factors, making their forecasting more difficult and requiring a more rigorous approach to integrate them into forecasting models.\u003c/p\u003e\n\u003cp\u003eIn conclusion, the results of the ADF and PP tests show that the calculated test statistics for all series are well below the critical values at the 1%, 5%, and 10% significance levels for the three models (without constant or trend, with constant, and with constant and trend). It is noteworthy that the CO2 emissions series is integrated of order 1. These results allow us to reject the null hypothesis of the presence of a unit root for all series, confirming the stationarity of the studied series.\u003c/p\u003e\n\u003cp id=\"_Toc164963681\"\u003e\u003cstrong\u003eTable 3: Estimations of the DCC Model (1.1) (Gold, CO2, and Temperature Anomaly)\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGold\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCo2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnomalies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel A : Mean Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003cp\u003e(0.567)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003cp\u003e(22.408) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.386\u003c/p\u003e\n \u003cp\u003e(56.886) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel B : Variance Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003cp\u003e(1.314)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(9.622) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eARCH(\u0026alpha;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e0.051\u003c/p\u003e\n \u003cp\u003e(2.462) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.170\u003c/p\u003e\n \u003cp\u003e(12.162) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.975\u003c/p\u003e\n \u003cp\u003e(21.722) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGARCH(\u0026beta;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e0.937\u003c/p\u003e\n \u003cp\u003e(34.606) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.828\u003c/p\u003e\n \u003cp\u003e(64.177) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0 (0.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026alpha;+\u0026beta;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e0.988\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e0.976\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel C : Dynamic Conditional Correlation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 396px;\"\u003e\n \u003cp\u003e0.117\u003c/p\u003e\n \u003cp\u003e(7.534) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 396px;\"\u003e\n \u003cp\u003e0.744\u003c/p\u003e\n \u003cp\u003e(18.866) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLog-Likelihood\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 396px;\"\u003e\n \u003cp\u003e11628.19\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel D : Residual Diagnostics Test\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 585px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eQ (20)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e36.433\u003c/p\u003e\n \u003cp\u003e(0.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e22039\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e22110\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 188px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e(*) Indicates statistical significance at the 5% level; Q (20) represents the autocorrelation test statistics applied to the residuals.\u003c/p\u003e\n\u003cp id=\"_Toc164963682\"\u003e\u003cstrong\u003eTable 4: Estimation of the DCC-GARCH Model (1.1) (Silver, CO2, and Temperature Anomaly)\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSilver\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCo2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnomalies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 567px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel A : Mean Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 567px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e-0.017 (-0.499)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e0.000 (22.408) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0.386 (56.886) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 567px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel B : Variance Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.017 (1.900)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e0 (0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0.001 (9.622) *\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eARCH(\u0026alpha;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.038 (5.662) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e0.170 (12.162) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0.975 (21.722) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGARCH(\u0026beta;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.956 (139.108) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e0.828 (64.177) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0 (0.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026alpha;+\u0026beta;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.994\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp\u003e0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0.976\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 567px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel C : Dynamic Conditional Correlation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 444px;\"\u003e\n \u003cp\u003e0.107 (8.330) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 444px;\"\u003e\n \u003cp\u003e0.766 (25.166) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLog-Likelihood\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e \u003ctd style=\"width: 188px;; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 444px;\"\u003e\n \u003cp\u003e10452.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 567px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel D : Diagnostic Tests on Residuals\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eQ (20)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e42.624\u003c/p\u003e\n \u003cp\u003e(0.002)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 146px;\"\u003e\n \u003cp\u003e22039\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003e22110\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e(*) Indicates the significance of values at the 5% threshold. Q (20): Represents the autocorrelation test statistics applied to the residuals.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 5. Estimation of the DCC-GARCH Model (1.1) (Platinum, CO2, and Temperature Anomaly)\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePlatinum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCo2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnomalies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel A : Mean Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e-0.013 (-0.398)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e0.004 (17.979) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e0.386 (56.658) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel B : Variance Equation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConstant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e0.013 (1.906)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e0 (0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e0.001 (9.663) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eARCH(\u0026alpha;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e0.036 (7.018) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e0.170 (12.162) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e0.975 (21.661) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGARCH(\u0026beta;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e0.959 (247.527) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e0.828 (64.177) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e0 (0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026alpha;+\u0026beta;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e0.996\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e0.976\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel C : Dynamic Conditional Correlation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 443px;\"\u003e\n \u003cp\u003e0.096 (7.616) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 443px;\"\u003e\n \u003cp\u003e0.787 (24.323) *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLog-Likelihood\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 188px; width: 20.0000%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 443px;\"\u003e\n \u003cp\u003e10431.74\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 443px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 565px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePanel D : Residual Diagnostic Tests\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eQ (20)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 137px;\"\u003e\n \u003cp\u003e75.268\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 151px;\"\u003e\n \u003cp\u003e22039\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 155px;\"\u003e\n \u003cp\u003e22110\u003c/p\u003e\n \u003cp\u003e(0.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;(*) Indicates the significance of values at the 5% threshold. Q (20): Represents the autocorrelation test statistics applied to the residuals.\u003c/p\u003e\n\u003cp\u003eThe analysis of the returns of gold, silver, and platinum through the DCC-GARCH (1,1) model highlights the predominant influence of environmental factors, such as CO2 emissions and temperature anomalies, on the dynamics of these precious metals. The results, presented in Tables 3, 4, and 5, show that, although the constants associated with the returns of gold (0.010), silver (-0.017), and platinum (-0.013) are not significant, the constants for CO2 (0.000 for gold and silver, 0.004 for platinum) and temperature anomalies (0.386 for all three metals) are highly significant, indicating their crucial role in the evolution of the mean dynamics. This aligns with the work of Karmakar et al. (2023), who also emphasize the importance of environmental factors in predicting the returns of precious metals.\u003c/p\u003e\n\u003cp\u003eIn terms of variance, the ARCH (\u0026alpha;) and GARCH (\u0026beta;) coefficients reveal a high persistence of volatility for the three precious metals studied. For gold, the coefficients are \u0026alpha; = 0.051 and \u0026beta; = 0.937; for silver, \u0026alpha; = 0.038 and \u0026beta; = 0.956; and for platinum, \u0026alpha; = 0.036 and \u0026beta; = 0.959. These results indicate that past shocks have a moderate but persistent effect on current volatility, which is crucial for long-term forecasting. For CO2, the coefficients \u0026alpha; = 0.170 and \u0026beta; = 0.828 also show volatility influenced by past shocks, while temperature anomalies present a very high \u0026alpha; (0.975) with no long-term persistence (\u0026beta; = 0). The sum of the coefficients \u0026alpha; + \u0026beta;, close to 1 for all variables (0.988 for gold, 0.994 for silver, 0.995 for platinum, 0.998 for CO2, and 0.976 for temperature anomalies), confirms a high persistence of volatility, essential for long-term forecasts.\u003c/p\u003e\n\u003cp\u003eIn terms of dynamic conditional correlation, the returns of these metals and the environmental factors show correlations influenced by past shocks and persistent over time, with coefficients A = 0.117 and B = 0.744 for gold, A = 0.107 and B = 0.766 for silver, and A = 0.121 and B = 0.781 for platinum. This complex dynamic justifies the use of the DCC-GARCH model to capture these interdependencies, which are crucial for investors and risk managers.\u003c/p\u003e\n\u003cp\u003eFinally, diagnostic tests on the residuals reveal significant autocorrelation, particularly for CO2 and temperature anomalies. This analysis highlights the importance of environmental factors in predicting the returns of precious metals. The persistence of volatility and the significant dynamic correlations between the returns and environmental variables make these factors essential for forecasting models. For investors, integrating these elements into their investment strategies is crucial in a context of growing environmental concerns, thus providing valuable insights for risk management and the development of sustainable investment strategies.\u003c/p\u003e\n\u003cp\u003eThe analysis of the accumulated impulse response functions, presented in Figure 1, reveals complex dynamics between carbon emissions, temperature anomalies, and precious metal returns before, during, and after the COVID-19 crisis.\u003c/p\u003e\n\u003cp\u003eFor gold, a CO2 shock before the pandemic led to a decrease in returns, illustrating a marked sensitivity to environmental pressures. However, after the pandemic, a reversal of this trend occurred, with an increase in returns, suggesting a possible adaptation or market response to a new post-crisis reality. Temperature anomalies generally had a negative effect on gold, with the notable exception of January 2, 2019, when a positive impact was observed. This particular date could correspond to exceptional circumstances or specific events that disrupted the usual relationship between temperature and the gold market.\u003c/p\u003e\n\u003cp\u003eRegarding platinum, CO2 shocks continued to have a negative impact on its returns throughout the studied periods, although volatility decreased after the crisis, suggesting some market stabilization. Responses to temperature anomalies followed a similar trajectory, with a marked negative return during the pandemic, followed by a positive recovery afterward, indicating potential resilience in the platinum market to climate disruptions.\u003c/p\u003e\n\u003cp\u003eSilver, on the other hand, exhibited particular resilience after the pandemic, with returns increasing in response to CO2 shocks, in contrast to other periods where the impact was negative. During the crisis, a temperature shock led to an increase in returns, suggesting a dynamic and potentially speculative response from investors to climate and health uncertainties.\u003c/p\u003e\n\u003cp\u003eThis analysis highlights the importance of environmental factors in determining precious metal returns, while underscoring the variability of responses depending on the economic and health periods. The recovery observed in the returns of gold and silver after the COVID-19 crisis may suggest that markets have adapted to new environmental risks, while platinum appears less responsive, with persistent negative returns despite reduced volatility. These findings open up interesting prospects for investment strategies that take into account evolving climate factors and their influence on financial assets. In line with the work of Miao et al. (2022), which emphasizes that external trends influence asset prices, these results suggest that the variability in the observed responses could be attributed to dynamics specific to each metal. These conclusions encourage further exploration of the predictive capacity of precious metal returns by more finely integrating carbon emissions and temperature anomalies into risk assessment models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 6. Forecasts from the TVC-VAR Model\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"447\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 124px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 323px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePerformance Measures\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRMSE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 112px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMAE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTheil\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGold\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e0.00880\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 112px;\"\u003e\n \u003cp\u003e0.0063\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.858\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSilver\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e0.0178\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 112px;\"\u003e\n \u003cp\u003e0.0129\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePlatinum\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e0.0181\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 112px;\"\u003e\n \u003cp\u003e0.0122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.889\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 6\u003c/strong\u003e highlights the performance of the TVC-VAR model for forecasting the returns of gold, silver, and platinum based on three key indicators: RMSE (Root Mean Square Error), MAE (Mean Absolute Error), and the Theil U coefficient.\u003c/p\u003e\n\u003cp\u003eFirst, gold stands out with an RMSE of 0.0088 and an MAE of 0.0063, suggesting strong model performance in forecasting its returns. The Theil U coefficient for gold is 0.858, indicating a good match between the forecasts and the actual values. This underscores the relative stability of gold, often regarded as a safe-haven asset during economic turbulence. This stability allows investors to have some confidence in the model\u0026apos;s predictions for gold, making it a strategic choice for those seeking predictable returns in an uncertain environment.\u003cbr\u003e\u0026nbsp;For silver, although the RMSE is slightly higher (0.0178) as well as the MAE (0.0129), the forecasts remain robust. The Theil U coefficient for silver shows that the model captures the dynamics of this metal well, although it is more sensitive to economic and industrial fluctuations. Due to its numerous industrial applications, silver often reacts to supply and demand shocks on a global scale. These results suggest that its returns can be anticipated with some degree of confidence, while keeping in mind that unexpected movements may occur. Investors might consider using these forecasts within a diversified strategy to mitigate the risks associated with silver.\u003c/p\u003e\n\u003cp\u003eRegarding platinum, the RMSE (0.0181) and MAE (0.0122) indicate that the model captures its fluctuations well, though the forecasts reveal slightly higher variability. The Theil U coefficient of 0.889 reflects the model\u0026apos;s ability to forecast platinum returns, even though this metal is influenced by specific factors such as environmental regulations and technological advances, particularly in the automotive industry. This means that investors interested in platinum may need to be more vigilant and adopt a more dynamic approach, considering external factors that could alter platinum\u0026apos;s returns in the medium and long term.\u003cbr\u003eThus, the results in \u003cstrong\u003eTable 6\u003c/strong\u003e show that the TVC-VAR model provides relevant forecasts for precious metals, with varying performances depending on the nature of the metal. Gold, with its stability, is particularly well-anticipated by the model, making it a strategic choice for investors seeking reliable forecasts. Silver and platinum, while presenting some volatility, are also satisfactorily modeled, but require more active management to leverage market dynamics. Finally, these results highlight the importance of environmental variables in forecasting precious metal returns, calling for increased consideration of these factors in investment decisions. These findings contrast with the work of Urquhart (2017), which identifies platinum as the most predictable, and Cohen (2022), who emphasizes the predictability of silver. This divergence suggests that accounting for environmental factors in modeling could alter the predictability order of precious metals and enrich the understanding of market dynamics.\u003c/p\u003e\n\u003cp\u003eIn this perspective, several recommendations emerge for market players in the precious metals sector, each tailored to the specific needs of the different groups involved.\u003cbr\u003e\u0026nbsp;For investors, it is crucial to diversify portfolios by including precious metals such as gold and silver, which offer relative stability and predictable returns in uncertain environments. However, the increased volatility of platinum requires a more nuanced approach, where investors must consider specific industrial and regulatory factors influencing this metal. For example, strict environmental regulations in the automotive industry could affect the demand for platinum, making its valuation more complex. The adoption of advanced econometric models, such as DCC-GARCH and TVC-VAR, can help investors better anticipate return fluctuations and manage associated risks. These models are particularly useful for capturing the effects of environmental shocks, such as CO2 emissions and temperature anomalies, which have a significant impact on precious metal markets. Furthermore, given the risks associated with negative asymmetry and leptokurtosis observed in returns, proactive risk management is essential. Investors can use derivative instruments, such as options or futures contracts, to protect themselves against extreme events and limit potential losses.\u003cbr\u003e\u0026nbsp;For policymakers and regulators, it is essential to implement public policies aimed at mitigating the impacts of climate change on financial markets. For example, measures to reduce CO2 emissions, such as carbon taxes or subsidies for green technologies, can have direct effects on the prices of precious metals. Moreover, raising awareness among market participants about climate risks and their influence on metal prices is crucial to promoting sustainable investment strategies. Regulators could also encourage academic research on the interactions between environmental factors and financial markets, for example, by funding interdisciplinary studies combining finance, economics, and climatology. These efforts would contribute not only to a better understanding of market dynamics but also to the development of regulatory frameworks adapted to current environmental challenges.\u003c/p\u003e\n\u003cp\u003eFor companies in the precious metals sector, it is crucial to adopt more sustainable production strategies to meet the growing demands of consumers and regulators. For instance, investing in clean technologies to reduce CO2 emissions related to metal extraction and refining could enhance their competitiveness in the market. Additionally, companies need to anticipate the impacts of climate anomalies on their supply chains and adapt their operations accordingly. For example, extreme weather events could disrupt mining activities, requiring robust continuity plans. Collaborating with researchers to explore sustainable alternatives to traditional precious metals could also open up new business opportunities.\u003cbr\u003e\u0026nbsp;Finally, for financial analysts and portfolio managers, it is essential to explicitly integrate environmental variables into their forecasting models. For example, including data on CO2 emissions and temperature anomalies in analyses could improve the accuracy of predictions and enable better anticipation of future trends. These professionals should also develop adaptive investment strategies that can quickly adjust to changes in climatic and regulatory conditions. For instance, overweighting assets less sensitive to environmental shocks or underweighting those exposed to strict regulations could optimize portfolio performance.\u003cbr\u003e\u0026nbsp;In sum, these recommendations aim to better address the risks and opportunities associated with precious metals in the context of ecological transition and increasing environmental regulations. They emphasize the importance of a proactive and interdisciplinary approach to navigating the complex challenges posed by the interactions between finance and the environment.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThe primary goal of this study was to examine the influence of climate variations and CO2 emissions on precious metal returns, specifically gold, silver, and platinum, while offering a dynamic and updated perspective compared to previous research. The results enrich our understanding of the complex relationships between environmental concerns and financial markets. They reveal that environmental factors profoundly alter the traditional perception of the predictability of precious metals. This analysis emphasizes the importance of integrating systemic risks related to climate change into investment strategies, opening new pathways for portfolio management and market regulation.\u003c/p\u003e \u003cp\u003eA complex dynamic emerges between environmental factors and the predictability of returns. The analysis of returns through the DCC-GARCH and TVC-VAR models highlights this relationship. Gold, with relatively stable volatility (α\u0026thinsp;=\u0026thinsp;0.051, β\u0026thinsp;=\u0026thinsp;0.937) and solid forecasts (RMSE\u0026thinsp;=\u0026thinsp;0.0088, MAE\u0026thinsp;=\u0026thinsp;0.0063), maintains its status as a reliable safe-haven asset. However, its growing sensitivity to CO2 emissions, as shown by the impulse response functions before and after the COVID-19 pandemic, indicates that environmental concerns may affect its traditional role as a safe-haven in the context of a climate crisis. In contrast, silver and platinum exhibit more volatile behaviors and require active management due to their high responsiveness to environmental and economic shocks. Platinum, in particular, shows increased persistence in its correlations with climate anomalies (A\u0026thinsp;=\u0026thinsp;0.096, B\u0026thinsp;=\u0026thinsp;0.787), while silver demonstrates post-pandemic resilience to CO2 shocks. These findings highlight that considering environmental factors is essential for anticipating market fluctuations in the context of ecological transition.\u003c/p\u003e \u003cp\u003eThe ARCH and GARCH coefficients reveal strong persistence of past shocks on current volatility for all three metals, confirming the importance of these factors for long-term forecasting. The sum of the coefficients α\u0026thinsp;+\u0026thinsp;β, close to 1 for all variables (0.988 for gold, 0.994 for silver, 0.995 for platinum, 0.998 for CO2, and 0.976 for temperature anomalies), illustrates strong inertia in volatility, crucial for long-term forecasts. The dynamic conditional correlations captured by the DCC-GARCH model reveal crucial interdependencies between precious metal returns and environmental variables. These interdependencies fully justify the use of advanced models to better understand the risks and opportunities associated with precious metals. Additionally, the accumulated impulse response functions show varied adjustments of precious metals to environmental shocks before, during, and after the COVID-19 pandemic, reflecting a progressive adaptation of markets to new economic and climate realities.\u003c/p\u003e \u003cp\u003eThe TVC-VAR model forecasts confirm these differentiated dynamics. Gold stands out for its relative stability and increased predictability, making it a strategic choice for investors seeking predictable returns. However, its sensitivity to CO2 emissions underscores the need to incorporate environmental concerns into investment strategies. Platinum, on the other hand, exhibits a stronger responsiveness to environmental shocks, requiring proactive management to mitigate risks associated with its dual role as an industrial and precious metal. Silver, although affected by similar factors, shows post-pandemic resilience and volatility linked to its industrial applications, which calls for a diversified approach to maximize returns.\u003c/p\u003e \u003cp\u003eIn conclusion, this study underscores the importance of integrating environmental concerns, such as CO2 emissions and climate anomalies, into the analysis of precious metal returns. It urges investors to adopt more dynamic and diversified approaches to better anticipate market fluctuations in the context of ecological transition. Regulators also play a key role in fostering policies that support the energy transition and the sustainable use of precious metals. Finally, these results open interesting avenues for future research on the impact of environmental policies and climate shocks on financial markets, as well as the integration of these factors into more robust economic models.\u003c/p\u003e \u003cp\u003eIn this context, several recommendations emerge for stakeholders in the precious metals market. For investors, it is recommended to diversify their portfolios with metals such as gold and silver while managing platinum's volatility using advanced models like DCC-GARCH and TVC-VAR. The use of derivatives is essential to mitigate risks associated with extreme events. Policymakers should promote policies that reduce CO2 emissions and raise awareness of the connection between climate and the metals market. Analysts and companies must integrate environmental factors into their models and strategies to anticipate regulatory and climate developments. These recommendations aim to better understand the risks and opportunities associated with precious metals in the context of ecological transition and increasing environmental regulations.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eIlyes Abidi: Was responsible for the empirical analysis, including the selection of methodology and econometric modeling. Ilyes developed the analytical framework, conducted the statistical analysis, and interpreted the results.Maissa Mejri: Contributed to data collection, the development of the theoretical framework, and the literature review. Maissa played a key role in gathering data and ensuring a solid theoretical foundation for the study.Mariem Nsaibi: Provided supervision and guidance throughout the research process. Mariem contributed to the overall direction and coherence of the manuscript, providing valuable editorial revisions to enhance clarity.Kamel Touhami: Contributed supervision, guidance, and editorial revisions to improve the manuscript\u0026rsquo;s clarity and coherence. Kamel provided insightful feedback to strengthen the quality of the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data used in this study were sourced from the following publicly available databases:\u0026bull; https://www.macrotrends.net/\u0026bull; https://gml.noaa.gov/\u0026bull; https://climatechange.umaine.edu/climate-matters/climate-reanalyzer/#:~:text=Climate%20Reanalyzer%20is%20a%20platform,data%20easily%20accessible%20by%20anyone.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbidi I, Touhami K (2024) Safe haven for crude oil: Bitcoin or precious metals? New insight from time varying coefficient-vector autoregressive model. Int J Energy Econ Policy 14(1):184\u0026ndash;195. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32479/ijeep.15107\u003c/span\u003e\u003cspan address=\"10.32479/ijeep.15107\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAddison T, Ghoshray A (2023) Discerning trends in international metal prices in the presence of nonstationary volatility. Resour Energy Econ 71:101334. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.reseneeco.2022.101334\u003c/span\u003e\u003cspan address=\"10.1016/j.reseneeco.2022.101334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaur DG, Lucey BM (2010) Is gold a hedge or a safe haven? An analysis of stocks, bonds and gold. Financial Rev 45(2):217\u0026ndash;229. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1540-6288.2010.00244.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1540-6288.2010.00244.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlackRock (2021) Investing with purpose 2021 annual report. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://s24.q4cdn.com/856567660\u003c/span\u003e\u003cspan address=\"https://s24.q4cdn.com/856567660\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCohen G (2022) Algorithmic strategies for precious metals price forecasting. Mathematics 10(7):1134. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/math10071134\u003c/span\u003e\u003cspan address=\"10.3390/math10071134\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDichtl H (2020) Forecasting excess returns of the gold market: Can we learn from stock market predictions? J Commod Markets 19:100106. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jcomm.2019.100106\u003c/span\u003e\u003cspan address=\"10.1016/j.jcomm.2019.100106\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDinh T, Goutte S, Nguyen DK, Walther T (2022) Economic drivers of volatility and correlation in precious metal markets. J Commod Markets 28:100242. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jcomm.2021.100242\u003c/span\u003e\u003cspan address=\"10.1016/j.jcomm.2021.100242\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEngle R (2002) Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models. J Bus Economic Stat 20(3):339\u0026ndash;350. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1198/073500102288618487\u003c/span\u003e\u003cspan address=\"10.1198/073500102288618487\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang J, Cao Y, Zhong P (2022) Searching for a safe haven to crude oil: Green bond or precious metals? Finance Res Lett 50:103303. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.frl.2022.103303\u003c/span\u003e\u003cspan address=\"10.1016/j.frl.2022.103303\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInternational Council on Mining and Metals (2019) Integrated mine closure. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.icmm.com/integrated-mine-closure\u003c/span\u003e\u003cspan address=\"https://www.icmm.com/integrated-mine-closure\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarmakar S, Gupta R, Cepni O, Rognone L (2023) Climate risks and predictability of the trading volume of gold: Evidence from an INGARCH model. Resour Policy 82:103438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.resourpol.2023.103438\u003c/span\u003e\u003cspan address=\"10.1016/j.resourpol.2023.103438\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMehrdoust F, Noorani M (2024) Forecasting precious metals price based on artificial neural network trained by L\u0026eacute;vy flight optimization algorithm. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2139/ssrn.4722754\u003c/span\u003e\u003cspan address=\"10.2139/ssrn.4722754\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. SSRN\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMetals Focus (2021) Reports. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.metalsfocus.com/reports/\u003c/span\u003e\u003cspan address=\"https://www.metalsfocus.com/reports/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiao M, Khaskheli A, Raza SA, Yousufi SQ (2022) Using internet search keyword data for predictability of precious metals prices: Evidence from non-parametric causality-in-quantiles approach. Resour Policy 75:102478. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.resourpol.2021.102478\u003c/span\u003e\u003cspan address=\"10.1016/j.resourpol.2021.102478\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNguyen D, Walther T (2020) Modeling and forecasting commodity market volatility with long-term economic and financial variables. J Forecast 39(2):126\u0026ndash;142. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/for.2616\u003c/span\u003e\u003cspan address=\"10.1002/for.2616\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapenfu\u0026szlig; P, Schischke A, Rathgeber A (2025) Factors of predictive power for metal commodities. North Am J Econ Finance 76:102309. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.najef.2024.102309\u003c/span\u003e\u003cspan address=\"10.1016/j.najef.2024.102309\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePierdzioch C, Risse M (2020) Forecasting precious metal returns with multivariate random forests. Empirical Economics 58:1167\u0026ndash;1184. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00181-018-1558-9\u003c/span\u003e\u003cspan address=\"10.1007/s00181-018-1558-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRaza SA, Masood A, Benkraiem R, Urom C (2023) Forecasting the volatility of precious metals prices with global economic policy uncertainty in pre and during the COVID-19 period: Novel evidence from the GARCH-MIDAS approach. Energy Econ 120:106591. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.eneco.2023.106591\u003c/span\u003e\u003cspan address=\"10.1016/j.eneco.2023.106591\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eResearch and Markets (2022) Global electric vehicle market analysis and forecast to 2027. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchandmarkets.com/\u003c/span\u003e\u003cspan address=\"https://www.researchandmarkets.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShahid MN, Jehanzeb M, Abbas A, Zubair A, Akbar MAH (2020) Predictability of precious metals and adaptive market hypothesis. Int J Emerg Markets 15(5):1011\u0026ndash;1027. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/IJOEM-07-2018-0404\u003c/span\u003e\u003cspan address=\"10.1108/IJOEM-07-2018-0404\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwamy PAVB, Hall SG, Tavlas GS, Hondroyiannis G (2010) Estimation of parameters in the presence of model misspecification and measurement error. Stud Nonlinear Dyn Econom 14(1):1\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2202/1558-3708.1743\u003c/span\u003e\u003cspan address=\"10.2202/1558-3708.1743\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUrquhart A (2017) How predictable are precious metal returns? Eur J Finance 23(14):1390\u0026ndash;1413. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/1351847X.2016.1204334\u003c/span\u003e\u003cspan address=\"10.1080/1351847X.2016.1204334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Gold Council (2020) Gold demand trends full year and Q4 2020. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-and-q4-2020/15507\u003c/span\u003e\u003cspan address=\"https://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-and-q4-2020/15507\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Gold Council (2022) Gold demand trends: Full year 2022. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-2022\u003c/span\u003e\u003cspan address=\"https://www.gold.org/goldhub/research/gold-demand-trends/gold-demand-trends-full-year-2022\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J, Agyei SK, Bossman A, Gubareva M, Marfo-Yiadom E (2024) Energy, metals, market uncertainties, and ESG stocks: Analysing predictability and safe havens. North Am J Econ Finance 69:102030. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.najef.2023.102030\u003c/span\u003e\u003cspan address=\"10.1016/j.najef.2023.102030\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y-J, Pan X (2021) Does the risk aversion of crude oil market investors have directional predictability for the precious metal and agricultural markets? China. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/CAER-08-2020-0275\u003c/span\u003e\u003cspan address=\"10.1108/CAER-08-2020-0275\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Agricultural Economic Review\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"CO₂, temperature anomalies, gold, silver, platinum, DCC-GARCH, TVC-VAR","lastPublishedDoi":"10.21203/rs.3.rs-6099418/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6099418/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study examines the influence of environmental factors, including CO₂ emissions and temperature anomalies, on the returns of precious metals (gold, silver, platinum) using DCC-GARCH and TVC-VAR models. The results show that gold exhibits relative stability with predictable returns, while silver and platinum, being more volatile, are subject to more pronounced fluctuations. CO₂ emissions and climate anomalies influence the volatility of precious metals by disrupting their supply and demand, while amplifying macroeconomic uncertainties and production costs, leading to increased persistence of past shocks on current volatility. The ARCH and GARCH coefficients reveal a strong persistence of these shocks for all metals, which is crucial for long-term forecasting. The dynamic conditional correlations captured by the DCC-GARCH model highlight critical interdependencies between the returns of precious metals and environmental variables. The accumulated impulse response functions show varying adjustments of precious metals to environmental shocks, reflecting a gradual adaptation of markets to new climate realities. Forecasts from the TVC-VAR model confirm the relevance of gold as a safe-haven asset, while silver and platinum require proactive management to mitigate the risks associated with their dual roles as both industrial and precious metals. These findings encourage investors, policymakers, and businesses to adopt sustainable and proactive strategies in the face of environmental challenges, while enriching the understanding of the complex interactions between finance and climate.\u003c/p\u003e","manuscriptTitle":"Can Environmental Factors Predict Precious Metal Returns?","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-11 06:48:04","doi":"10.21203/rs.3.rs-6099418/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4e640930-4ac5-4e36-9522-ce4e79bcb9bd","owner":[],"postedDate":"March 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":45356166,"name":"Business and commerce/Economics"},{"id":45356167,"name":"Business and commerce/Finance"},{"id":45356168,"name":"Social science/Environmental studies"}],"tags":[],"updatedAt":"2025-03-11T06:48:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-11 06:48:04","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6099418","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6099418","identity":"rs-6099418","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}