VOLATILITY SPILLOVERS AND HEDGING EFFECTIVENESS OF STRATEGIC
AND PRECIOUS METALS IN CLEAN ENERGY, EV, AND AUTOMOTIVE EQUITY
MARKETS: EVIDENCE FROM DCC-GARCH, AND GRANGER CAUSALITY
MODELS
by
Soleiman Hashemishahraki
M.Sc., University of Shahid Beheshti, 2015
B.Sc., University of Mazandaran, 2012

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
BUSINESS ADMINISTRATION

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
July 2025
© Soleiman Hashemishahraki, 2025

Declaration of Originality and Data Authenticity
I hereby declare that this thesis is the result of my own original work and has not been
submitted for a degree at any other institution. All sources of information and data used in this
study have been properly acknowledged and referenced. I also confirm that the data used in
this research are authentic, have been collected from reliable sources, and have not been
fabricated or manipulated in any way.
This work was completed under the direct supervision of Dr. Chengbo Fu at the
University of Northern British Columbia (UNBC).
Soleiman Hashemishahraki
Business Administration
University of Northern British Columbia
24 Jun 2025

ii

Abstract
This thesis explores the dynamic relationships between strategic and precious metals
and equity markets, including electric vehicle (EV) manufacturers, traditional automakers, and
clean energy indices. Using daily data from 2014 to 2024, the study applies a range of
econometric techniques—DCC-GARCH, VAR, Granger causality tests, impulse response
functions (IRFs), and the Diebold and Yilmaz (2012) spillover index, to analyze volatility
transmission, co-movement, hedge ratios, and optimal portfolio strategies.
The results show strong time-varying correlations and spillovers between strategic
metals (such as copper, aluminum, cobalt, nickel, and zinc) and EV or clean energy stocks,
particularly during major crises like the COVID-19 pandemic and the Russia–Ukraine war.
Granger causality and IRF analyses reveal that strategic metals have a more significant impact
on EV stocks than on traditional automakers, reflecting their critical role in battery and
renewable technology supply chains.
Hedging and portfolio analysis further indicate that strategic metals offer more
effective, responsive hedging tools for green investments, while precious metals like gold and
silver provide more passive protection during market uncertainty. Platinum is found to be a
relatively inefficient hedge due to its high volatility and cost.
Overall, the study contributes to the literature by demonstrating the growing financial
integration of strategic metals with the green economy and offering actionable insights for
investors and policymakers seeking to manage risk and enhance portfolio resilience during
economic and geopolitical shocks.
Keywords: Strategic Metals, Precious Metals, Electric Vehicles, Clean Energy Stocks,
DCC-GARCH, Hedging, Volatility Spillovers
iii

Table of Contents
Declaration of Originality and Data Authenticity ....................................................... ii
Abstract ................................................................................................................. iii
List of Tables .......................................................................................................... vi
List of Figures ........................................................................................................ vi
Acknowledgments .................................................................................................. vii
Chapter One: INTRODUCTION .............................................................................. 1
1. 1

Research Background and Motivation ............................................................ 1

1. 2

Research Problem ........................................................................................ 5

1. 3

Objectives of the Study ................................................................................ 6

1. 4

Research Contribution .................................................................................. 8

1. 5

Research Questions and Hypotheses ............................................................... 9

1. 6

Structure of the Thesis ................................................................................ 10

Chapter Two: LITERATURE REVIEW.................................................................. 11
2. 1

Financialization of Strategic and Precious Metals ........................................... 12

2. 2

Metal-Equity Linkages: Correlation and Volatility .......................................... 14

2.2.1

Strategic (Non-Precious) Metals and Equity Market Dynamics .................. 14

2.2.2
Precious Metals and Equity Markets: Hedging, Safe-Haven, and Volatility
Spillovers ........................................................................................................ 17
2. 3

Hedging Effectiveness of Metals in Equity Portfolios...................................... 20

2. 4

Methodologies in the Literature: DCC-GARCH, VAR, Granger Causality.......... 22

2.4.1

Dynamic Conditional Correlation GARCH (DCC-GARCH) ..................... 22

2.4.2

Vector Autoregression (VAR) ............................................................... 23

2.4.3

Granger Causality ............................................................................... 25

2.4.4

Total Connectedness Index (TCI) .......................................................... 26

2.4.5

Impulse Response Functions (IRF) ........................................................ 27

2. 5

Gaps in the Literature and Research Justification ............................................ 29

Chapter Three: RESEARCH METHODOLOGY .................................................... 31
3.1

Research Design ........................................................................................ 31

3.2

Data Collection and Variable Selection ......................................................... 32
iv

3.2.1

Data Sources and Variable Definitions ................................................... 32

3.2.2

Data Processing .................................................................................. 33

3.2.3

Rationale for Variable Selection ............................................................ 34

3.3

Model Specifications ................................................................................. 35

3.3.1

Dynamic Conditional Correlation GARCH (DCC-GARCH) ..................... 35

3.3.2

Vector Autoregression (VAR) ............................................................... 38

3.3.3

Granger Causality Tests ....................................................................... 40

3.3.4

Impulse Response Functions (IRF) ........................................................ 41

3.3.5

Diebold and Yilmaz (2012) Approach .................................................... 42

3.4

Limitations and justifications of Selected Methodologies ................................. 46

Chapter Four: EMPIRICAL RESULTS .................................................................. 47
4.1

Main results .............................................................................................. 48

4.2

Volatility Connectedness Results (Diebold and Yilmaz, 2012) .......................... 57

4.3

Dynamic Conditional Correlation GARCH (DCC-GARCH) Results ................. 63

4.4

Granger Causality Results ........................................................................... 82

4.5

Impulse Response Functions (IRF) Results .................................................... 91

4.6

Hedging Analysis and Portfolio Weights ....................................................... 97

4.7

Robustness of Empirical Findings .............................................................. 114

Chapter Five: CONCLUSION .............................................................................. 115
BIBLIOGRAPHY ............................................................................................... 117

v

List of Tables
Table 1: Descriptive Statistics ................................................................................... 51
Table 2: Dynamic Conditional Correlation (DCC-GARCH) Results for Automakers, EV
Manufacturers, and Clean Energy Indices Paired with Metals ........................................ 65
Table 3 Statistical Summary of Dynamic Conditional Correlations Between equities and
metals .................................................................................................................... 79
Table 4 Diagnostic test ............................................................................................. 81
Table 5 Granger Causality Test Results Between Strategic/Precious Metals and Equity
Markets (Bidirectional Analysis) ............................................................................... 85
Table 6 summary statistics for the optimal hedge ratios across different equity indices and
metals .................................................................................................................. 103
Table 7 Summary statistics for the optimal weights across different equity indices and metals
........................................................................................................................... 112

List of Figures
Fig. 1 Correlation Heatmap of Strategic Metals, Precious Metals, EV Manufacturers,
Traditional Automakers, and Clean Energy Indices ....................................................... 52
Fig. 2 Rolling Correlation between Strategic Metals and Equity Sectors (Automakers, EV
Stocks, Clean Energy, and Precious Metals) ................................................................ 55
Fig. 3 Static Volatility Spillovers Among Strategic Metals, Automakers, Clean Energy, and
Precious Metals (2014–2024) .................................................................................... 58
Fig. 4 Total Connectedness Index (TCI) from Jan 30, 2014, to Jan 30, 2024 ..................... 62
Fig. 5 Comparative Time Series of Rolling Conditional Correlations and Covariances
between Equity Sectors and Metal Markets. ................................................................ 76
Fig. 6 Granger Causality Directions Between Metals and Equities .................................. 91
Fig. 7 Impulse Response Functions (IRFs): Metal Shocks to Equity Reactions ................. 97
Fig. 8 Time‐varying hedge ratios between equity indices and metals estimated via DCCGARCH models. ................................................................................................... 102
Fig. 9 Boxplots of Time-Varying Hedge Ratios Between Equity Stocks and
Strategic/Precious Metals Across Crisis and Normal Periods. ....................................... 111
Fig. 10 Robustness check based on total spillovers from Diebold and Yilmaz (2012). RW,
rolling window. ..................................................................................................... 115

vi

Acknowledgments
I would like to express my sincere gratitude to my supervisor, Dr. Chengbo Fu, for his
invaluable guidance, encouragement, and unwavering support throughout my thesis journey.
His expertise and insightful feedback were instrumental in shaping the direction and quality of
this research. I am especially thankful for the opportunity to work as a research assistant under
his supervision, which greatly enriched my academic and professional development.
I would also like to thank my committee members, Dr. Jing Chen and Dr. Fan Jiang,
for their support and constructive input throughout this study.
My heartfelt thanks go to all the faculty members of the Business Administration
Department at the University of Northern British Columbia. Their teaching and mentorship
have played a significant role in my academic growth.
Finally, I am deeply grateful to my parents for their unconditional love, patience, and
sacrifices. Their unwavering belief in me has been my greatest source of strength and
inspiration.

vii

Chapter One: INTRODUCTION
1. 1 Research Background and Motivation
The increased use of electric vehicles (EVs), wind power equipment, solar panel
installations, and energy storage equipment has led to an unprecedented increase in the demand
for essential metals such as lithium, nickel, cobalt, copper, and rare earth elements. The
International Energy Agency (IEA, 2021) reports that achieving the Paris Agreement targets
can increase the world's demand for these metals six times by 2040. The demand increase is
primarily driven by the widespread use of EVs and clean technology, in which the metals are
used in battery manufacturing, electrification, and clean infrastructure.
This change is not only reconfiguring global markets for commodities but also causing
concern regarding price volatility, security of resources, and the geopolitical dimensions of
mineral extraction. (Gustafsson et al., 2022) state that metals such as cobalt, nickel, and lithium
are being pulled into financial markets. Hence, they are subject to the investor's sentiment and
speculative trading, particularly in the portfolios of clean energy investments. Their dual role,
as industrial inputs and speculative financial assets, makes them increasingly sensitive to
macro-financial dynamics.
As a result, the volatility of these metals is no longer purely industrial; it also
contributes to uncertainty in financial markets. This presents challenges and opportunities for
investors and policymakers in managing risk and building resilient, sustainability-oriented
portfolios. Recognizing and responding to the dynamic financial behavior of strategic metals
is becoming crucial in shaping effective strategies for resource security and market stability
amid the ongoing energy transition.

1

Beyond their industrial roles, strategic metals have increasingly become financial
assets and are no longer viewed solely as commodities. Their growing relevance in global
financial systems has drawn increased attention from institutional investors and portfolio
managers. As noted by (Cagli, 2023), battery metals demonstrate significant return and
volatility transmission with the equities of future mobility firms, signaling their active
integration into financial systems. Consequently, understanding how these metals interact with
other asset classes is now an essential dimension of portfolio design and risk hedging.
The relationship between strategic metals and clean energy equities, especially electric
vehicle manufacturers, has become a central research focus. Studies by (Alekseev et al., 2024;
Y. Chen et al., 2022) show that price movements in metals like cobalt and nickel directly impact
the valuation of firms along the EV supply chain. These metals form a significant part of input
costs, and their volatility transmits to equity markets through various channels such as
operational margins, supply disruptions, and investor expectations. Moreover, this comovement tends to strengthen during macroeconomic stress or geopolitical instability.
Empirical evidence supports the presence of volatility spillovers and time-varying
correlations between metal prices and clean energy financial instruments. For instance, Cagli
(2023) demonstrates that strategic metals often transmit volatility to companies focused on
future mobility through frequency-based methods. Similarly, Bastianin et al. (2025) report that
volatility in transition metals has increased significantly over the past decade, especially during
financial crises. These findings reinforce the importance of modeling these relationships
dynamically to understand better the risks embedded in clean energy portfolios.
Recent literature also explores the potential for strategic and precious metals to act as
hedging instruments or safe-haven assets. Gold and silver, long considered traditional hedges,
2

have been shown to protect against equity downturns (Elie et al., 2019; Erdoğan et al., 2022).
More recent studies by Gustafsson et al. (2022) and Rajwani et al. (2023) suggest that industrial
metals like cobalt and nickel may offer partial hedging benefits during heightened market
uncertainty. However, their effectiveness appears to vary depending on the broader volatility
regime.
Understanding the dynamics of interactions between strategic and precious metals and
equity markets is of scholarly interest and highly pertinent to investors aiming to maximize
portfolio performance. The search for efficient hedging tools is increasingly crucial,
considering the increasing inclusion of clean energy and EV-related assets into investor
portfolios. Dynamic correlation models like the DCC-GARCH enable investors to estimate
time-varying hedge ratios and deal with risk exposure more effectively. Research by Pakrooh
and Manera (2024) and Rajwani et al. (2023) indicates that including energy transition metals
in equity portfolios improves risk-adjusted returns and reduces vulnerability to tail risks,
especially when correlations are regularly updated in response to market changes. Furthermore,
the distinct behavior of metals during macroeconomic and geopolitical shocks offers
diversification advantages.
Despite growing academic interest in the financial behavior of both strategic and
precious metals, the literature still lacks a comprehensive comparative analysis of their roles
across clean energy, EV, and traditional automotive sectors. Most existing studies examine
either group of metals in isolation or focus narrowly on one equity market segment. A broader,
more integrated perspective is still missing, particularly one considering multiple metals and
stock sectors.

3

This limited scope has left important questions unanswered. For example, do strategic
metals better hedge EV stocks than precious metals? How are different classes of assets
influenced by spillovers of volatility? And to what degree are these dynamics influenced by
external shocks such as the COVID-19 pandemic or the Russia-Ukraine conflict? These
questions necessitate the application of strong, time-varying econometric models that can
capture asymmetric behavior, crisis, and asset-implied relations. Few studies have employed
an integrated framework using DCC-GARCH, VAR, Granger causality, and connectedness
analysis.
The global shocks induced by COVID-19 and the Russia–Ukraine war have
underscored the importance of crisis-aware financial modeling. Both events triggered extreme
volatility in commodity and equity markets, disrupting metal supply chains and amplifying
their spillover effects on financial assets. As noted by (Cagli, 2023; Ghosh et al., 2023),
volatility transmission among asset classes is known to become more acute in crises, thus
emphasizing dynamic hedging techniques and adaptive risk management measures.
This thesis endeavors to bridge this gap in the literature by comprehensively analyzing
the behavior of strategic and precious metals in relation to sectoral equity indices, clean energy,
EV manufacturers, and traditional automakers, under both normal and crisis conditions. It
applies a combination of econometric models to capture the evolving relationships between
these assets and offers empirical insights for risk management, policy planning, and investment
decision-making.
Ultimately, this study contributes to the growing literature on financialization in the
clean energy era by examining multiple metals and stock sectors. Its findings are intended to

4

support investors, analysts, and policymakers in designing strategies that respond to the
financial risks and opportunities of a global shift toward sustainability.
1. 2 Research Problem
The increasing integration of clean energy technologies, electric vehicles (EVs), and
sustainability-focused investments has elevated the importance of strategic metals as industrial
inputs and as influential financial instruments. Metals such as cobalt, nickel, copper, and
lithium are now tightly linked to the economic performance of EV manufacturers and clean
energy indices. However, while literature acknowledges these links, a significant analytical
gap remains.
Existing research treats strategic and precious metals separately, often evaluating their
hedging or volatility transmission capabilities in isolation or focusing on a limited set of clean
energy equities. For instance, while studies by (Alekseev et al., 2024; Y. Chen et al., 2022)
document the influence of battery metals on EV stocks, they typically do not compare their
behavior with traditional safe-haven assets like gold and silver, nor do they systematically
evaluate sector-specific differences between EVs, clean energy indices, and traditional
automakers. The absence of a unified, comparative framework limits our understanding of how
metals behave across different market segments and conditions.
Moreover, many existing studies rely on static or limited econometric models with
constant relationships between metal and equity markets. The models do not reflect timevarying correlations, asymmetric spillovers, or time-varying causal patterns, which are crucial
for financial instruments under real-market conditions. Events such as the COVID-19
pandemic and the Russia–Ukraine conflict have revealed how rapidly these relationships can
change. As (Ghosh et al., 2023) highlighted, volatility transmission between energy and metals
5

intensified significantly during crisis periods, and market connectedness became more
complex and nonlinear.
Despite these developments, few studies have employed an integrated econometric
framework combining dynamic conditional correlation (DCC-GARCH), vector autoregression
(VAR), Granger causality, and connectedness analysis to examine the evolving behavior of
strategic and precious metals across multiple equity sectors. This lack of comprehensive
modeling limits our understanding of how metals respond to market stress, whether they act as
volatility transmitters or risk reducers, and how their role varies between clean energy, EV, and
traditional auto sectors.
This thesis addresses these gaps by adopting a unified and dynamic econometric
framework that integrates DCC-GARCH, VAR, Granger causality testing, and total
connectedness index (TCI) analysis. These models allow for exploring time-varying
correlations, directional volatility spillovers, and predictive causality across different asset
classes. The study covers a broad set of strategic and precious metals and compares their
behavior across clean energy indices, EV stocks, and traditional automakers under normal and
crisis conditions. By doing so, the study aims to determine whether strategic metals serve as
more effective hedging instruments than precious metals, whether they transmit or absorb
volatility during shocks, and how their financial role shifts in response to systemic stress. The
findings will provide practical insights for investors, portfolio managers, and policymakers
seeking to build more resilient, sustainability-oriented financial strategies.
1. 3 Objectives of the Study
The main objective of this research is to examine the dynamic financial relationships
and hedging effectiveness of both strategic and precious metals in relation to electric vehicle

6

(EV) stocks, traditional automakers, and clean energy indices. By employing advanced
econometric models, the study aims to explore the dynamic behavior of these asset classes
under financial uncertainty, particularly during periods of crisis, such as the COVID-19
pandemic and the Russia–Ukraine war.
The specific objectives of this study are as follows:
•

To evaluate the dynamic correlations between strategic metals, precious metals, and the
returns of EV stocks, traditional automakers, and clean energy indices using the DCCGARCH model.

•

To assess the volatility transmission and predictive relationships between strategic
metals, precious metals, and stock returns through Vector Autoregressive (VAR)
models and Granger causality tests.

•

To compute and compare dynamic hedge ratios between strategic metals and precious
metals and EV/traditional automaker stocks, identifying the effectiveness of hedging
over time.

•

To analyze the hedging effectiveness of strategic metals relative to traditional precious
metals (e.g., gold, silver) across EV and non-EV (automaker) portfolios.

•

To estimate optimal portfolio weights involving strategic and precious metals and
assess their ability to reduce risk in clean energy and automotive investment portfolios.

To investigate the impact of external shocks, including the COVID-19 pandemic and the
Russia–Ukraine war, on the degree of connectedness between strategic metals and EV-related
financial assets using total connectedness index (TCI) analysis.

7

1. 4 Research Contribution
This research makes several contributions to the growing body of literature on the
financial behavior of strategic and precious metals in relation to clean energy and mobilityfocused equity markets. While prior studies have examined individual relationships between
selected metals and clean energy stocks, few have offered a comprehensive, comparative
analysis that spans multiple sectors, including electric vehicle manufacturers, clean energy
indices, traditional automakers, and various metals.
Firstly, this study contributes by integrating strategic metals (e.g., cobalt, nickel,
copper, lithium) and traditional precious metals (e.g., gold, silver, platinum, palladium) into a
unified empirical framework. This integration allows for a side-by-side assessment of their
hedging performance, volatility transmission behavior, and portfolio relevance, which remains
underexplored mainly in existing research.
Secondly, the study introduces a multi-method econometric approach by combining
dynamic conditional correlation (DCC-GARCH), vector autoregression (VAR), Granger
causality analysis, and total connectedness index (TCI) modeling. This comprehensive
framework captures time-varying correlations, spillover intensity, and predictive causality in a
way that reflects real-world market dynamics, especially under crisis conditions such as the
COVID-19 pandemic and the Russia–Ukraine war. Few studies to date have applied all of these
models in tandem across such a diverse asset group.
Thirdly, this research expands the literature by providing a sector-level comparison of
hedge effectiveness and risk dynamics across clean energy, EV, and traditional automotive
stocks. In doing so, it sheds light on whether strategic metals offer superior hedging capabilities
relative to precious metals, particularly during periods of elevated market uncertainty. The

8

results are expected to interest scholars in energy finance and sustainability investing, as well
as institutional investors and policymakers focused on portfolio resilience, resource risk
management, and green investment planning.
Ultimately, this thesis offers fresh empirical insights into how commodities and
financial markets interact during the clean energy transition. These insights can help enhance
theoretical models and create practical risk management strategies in economies prioritizing
sustainability.
1. 5 Research Questions and Hypotheses
Based on the background and objectives outlined in the previous sections, this study
aims to answer the following research questions:
•

Q1: Do strategic metals exhibit significant time-varying correlations with clean energy
indices, electric vehicle (EV) stocks, and traditional automaker stocks?

•

Q2: Do strategic metals substantially influence EV stocks' return volatility more than
traditional automaker stocks?

•

Q3: Does strategic metals Granger cause the returns of EV companies to be more
significant than those of traditional automakers?

•

Q4: Are hedge ratios involving strategic metals more effective for EV stocks than
traditional automaker stocks?

•

Q5: Do strategic metals outperform precious metals in hedging exposure to EV stocks?

•

Q6: Do optimal portfolio weights involving strategic metals reduce portfolio risk more
effectively in EV-focused portfolios than in those focused on traditional automakers?

9

•

Q7: Have external shocks, such as the COVID-19 pandemic and the Russia–Ukraine
conflict, intensified the volatility connectedness between strategic metals and EVrelated financial assets?
These questions give rise to the following testable hypotheses:

•

H1: A significant time-varying correlation exists between strategic metals and clean
energy indices, EV stocks, and traditional automaker stocks.

•

H2: Strategic metals (e.g., cobalt, nickel, copper, aluminum) have a greater influence
on the return volatility of EV stocks than on traditional automaker stocks.

•

H3: Strategic metals Granger-cause EV companies' returns are more significant than
traditional automakers.

•

H4: Strategic metals provide more effective hedge ratios for EV stocks than traditional
automaker stocks.

•

H5: Strategic metals offer greater hedging effectiveness for EV stocks than precious
metals such as gold and silver.

•

H6: Optimal portfolio weights that include strategic metals reduce portfolio risk more
effectively in EV-focused portfolios than traditional automaker portfolios.

•

H7: The COVID-19 pandemic and geopolitical shocks (e.g., the Russia–Ukraine war)
increased the volatility and connectedness between strategic metals and EV-related
financial assets.

1. 6 Structure of the Thesis
The remainder of this thesis is organized as follows:

10

•

Chapter Two reviews the relevant academic literature on the interlinkages between
strategic and precious metals, energy-related financial markets, and volatility
modeling techniques.

•

Chapter Three details the research methodology, including data sources, variable
selection, model specifications, and diagnostic tests.

•

Chapter Four presents empirical results, including dynamic correlation patterns,
hedge ratio analysis, and VAR-based causality outcomes.

•

Chapter Five concludes the thesis by summarizing the main findings, discussing
their practical implications, and proposing avenues for future research.
Chapter Two: LITERATURE REVIEW

This chapter reviews the existing academic research on the financial behavior of
strategic and precious metals, particularly in relation to clean energy and electric vehicle (EV)
markets. It highlights how non-precious metals, such as lithium, cobalt, nickel, and copper,
have evolved from being viewed purely as industrial inputs to being recognized as influential
financial assets. Their increasing relevance in investment strategies has drawn interest for their
potential roles in hedging, portfolio diversification, and managing exposure to market
volatility.
By focusing on dynamic financial relationships, such as volatility spillovers, timevarying correlations, and systemic connectedness during periods of crisis, this chapter sets the
stage for identifying the empirical gaps that this study aims to address. The literature review is

11

organized around four key themes, each reflecting a core area of investigation and helping to
identify the research gaps that this study aims to address.
2. 1 Financialization of Strategic and Precious Metals
In recent years, the role of both strategic and precious metals has undergone a
significant transformation, extending far beyond their traditional industrial or monetary
functions. Once valued mainly for manufacturing components like lithium, nickel, cobalt, and
copper or their safe-haven appeal like gold, silver, palladium, and platinum, these metals have
now stepped firmly into the world of financial assets. They're increasingly responsive to shifts
in investor sentiment, global risk appetite, and broader market conditions (Cagli, 2023; Kang
et al., 2023). This transition is part of a larger phenomenon known as the financialization of
commodity markets, where the pricing of metals is no longer solely determined by the
fundamental forces of supply and demand but increasingly reflects the dynamics of financial
markets(Chang et al., 2017; Westra, 2024).
This transition is part of a broader phenomenon often referred to as the financialization
of commodity markets. In this context, metals are no longer priced solely based on supplydemand fundamentals but increasingly reflect financial market dynamics. Institutional
investors and significant funds have introduced new layers of volatility, trading metals via
exchange-traded funds (ETFs), futures, and commodity index products. As highlighted by
(Włodarczyk & Szturo, 2018), financialization has fundamentally changed the structure of
commodity markets by integrating them more closely with global capital flows.
Strategic metals such as lithium, cobalt, nickel, and copper have gained particular
relevance in this shift. These metals are essential components of clean technologies like electric
vehicle batteries and solar panels, which have attracted long-term investor interest. However,

12

their prices are now influenced by industrial growth and their perceived financial value. (Yin
& Cao, 2024) note that this dual role exposes these metals to short-term speculation and longhorizon portfolio strategies.
Precious metals have followed a similar trajectory, albeit originating from a different
historical context. Gold and silver have long been considered safe-haven assets and stores of
value, particularly during periods of market stress (Baur & Lucey, 2010a). However, their
integration into structured financial products, such as commodity funds and ETF-based
portfolios, has strengthened their correlation with equity markets. (Jégourel, 2021) highlights
how metals once viewed as purely industrial or monetary now appear in volatility hedging
models, risk management strategies, and ESG-focused investment vehicles.
The implications of this shift are significant. Ding et al. (2021) show that financialized
metals exhibit greater price fluctuations and stronger co-movements with equity markets,
especially during periods of macroeconomic stress. This makes them both a potential source
of portfolio diversification and a channel of contagion during crises. Futures markets, once
used primarily by producers and industrial users for hedging, are now dominated by
speculative flows that drive short-term volatility (Mayer et al., 2017). In effect, metals have
become mirrors of broader financial uncertainty.
These evolving financial characteristics suggest that strategic and precious metals are
no longer isolated commodities. Instead, they are deeply interconnected with broader financial
systems, investor expectations, and systemic market risk. These evolving characteristics justify

13

the application of dynamic econometric models, such as DCC-GARCH and VAR, better
equipped to capture non-linear, time-varying relationships between metals and equity markets.
2. 2 Metal-Equity Linkages: Correlation and Volatility
In recent years, metals have become increasingly linked to equity markets, exhibiting
dynamic behaviors such as time-varying correlations, volatility spillovers, and systemic
connectedness. These financial interactions are especially pronounced in strategic metals
associated with clean technologies and in precious metals traditionally viewed as safe-haven
assets (Banerjee and Pradhan, 2024; Chen et al., 2022). To better understand these differences,
this section is divided into two parts: strategic metals and precious metals, each analyzed in
terms of their correlation dynamics, volatility behavior, and roles in financial portfolios.
2.2.1

Strategic (Non-Precious) Metals and Equity Market Dynamics

Strategic metals such as lithium, cobalt, nickel, and copper have become increasingly
integral to the global shift towards clean energy and electrification. Beyond their industrial
applications, these metals have emerged as significant financial assets, exhibiting dynamic
interactions with equity markets, particularly in sectors like electric vehicles (EVs) and
renewable energy (Bastianin et al., 2025; IEA, 2021).
(Y. Chen et al., 2022) explored the spillover effects and hedging effectiveness between
non-ferrous metals and sub-sectoral clean energy stocks. Utilizing time and frequency domain
analyses, they found that spillovers are predominantly short-term and intensify during periods
of crisis. Notably, non-ferrous metals acted as effective hedging instruments, especially for
developer stocks within the clean energy sector.
Alekseev et al. (2024) investigated the return and volatility spillovers between raw
material markets for EV batteries and downstream EV producers. Employing an EGARCH

14

model, their study revealed significant return spillovers from lithium producers to EV
manufacturers, indicating a strong linkage between upstream material markets and
downstream equity performance.
Cagli (2023) examined the volatility spillover between battery metals and future
mobility stocks using a time-varying frequency connectedness approach. The study highlighted
that volatility transmission is more pronounced during periods of market stress, emphasizing
the interconnectedness between battery metal prices and mobility-related equities.
Ghosh et al. (2023)analyzed the quantile connectedness between energy and metal
markets during the COVID-19 pandemic. Their findings indicated a sharp increase in
connectedness post-pandemic, with stronger dependencies at higher quantiles. Among energy
metals, cobalt exhibited the least connection to energy markets, suggesting varying degrees of
integration among different metals.
Bastianin et al. (2025)focused on forecasting the volatility of energy transition metals
(ETMs), including copper, lithium, nickel, and cobalt. Their comprehensive analysis identified
significant heterogeneity in ETM volatility patterns, challenging standard classifications and
underscoring the complexity of modeling these markets.
Pakrooh and Manera (2024) explored causality, connectedness, and volatility passthrough among energy, metal, stock, and carbon markets within the European Union. Their
study demonstrated that the EU carbon market is a net receiver of shocks from energy, metal,
and financial markets, highlighting the systemic importance of strategic metals in broader
market dynamics.
Therefore, these studies underscore the evolving financialization of strategic metals
and their significant interactions with equity markets. Understanding these dynamics is crucial

15

for investors and policymakers aiming to navigate the complexities of modern financial
systems intertwined with the clean energy transition.
Recent scholarship by Fu and colleagues has significantly expanded the understanding
of how political uncertainty, ESG disclosures, and macro-financial risks affect energy markets
and commodity-linked firms. Gong et al. (2022) examine how international political
uncertainty, proxied by U.S. presidential elections—interacts with firm-level climate exposure
to amplify stock return volatility across 34 countries. Extending this line of inquiry, Gong et
al. (2023) show that fossil fuel companies exhibit heightened sensitivity to climate risk,
especially in more developed financial markets. Using a Markov switching VAR model, Gong
et al. (2021) identify distinct oil price regimes and their spillover effects on firm behavior in
the global energy sector.
In the context of energy policy and valuation, Gong and Jia (2024) investigate the
impact of U.S. political decisions during the Paris Agreement era, finding asymmetric
valuation effects between renewable and fossil energy firms. Similarly, Clancey-Shang and Fu
(2024) focus on corporate social responsibility (CSR) disclosure amid the Russia–Ukraine
conflict, revealing that strong ESG commitments can enhance market quality during
geopolitical shocks. In a related study, Clancey-Shang and Fu (2023) compare the performance
of foreign and domestic stocks on U.S. markets during the same conflict, highlighting how
political risk reshapes cross-border investment behavior.
Other contributions emphasize labor market and firm-level dynamics. Liu et al. (2024)
explore how the 2014–2015 oil price collapse influenced labor investment among Chinese
firms, documenting substantial real-sector adjustment during commodity shocks. Li et al.
(2025) provide a comparative analysis of traditional and green energy firms based on ESG

16

activity, free cash flow, and market valuation, underscoring ESG’s rising role in investor
decision-making. Fu et al. (2025a) analyze mortgage credit and sea-level rise (SLR) risk in
U.S. housing markets, showing how environmental exposure affects credit access. Lastly, Fu
et al. (2025b)examine the value of investor sophistication across capital markets, emphasizing
the growing importance of behavioral factors in pricing risk and opportunity in ESG-focused
portfolios.
2.2.2

Precious Metals and Equity Markets: Hedging, Safe-Haven, and Volatility
Spillovers

Precious metals, such as gold, silver, platinum, and palladium, have long been viewed
as safe havens during times of market instability. Their scarcity, high liquidity, and generally
low correlation with equities make them valuable tools for hedging against market
downturns(Baur & McDermott, 2010). In recent years, however, their role has expanded. As
precious metals become more integrated into investment portfolios and financial products, they
are increasingly used for diversification and risk management (Reboredo, 2013). This has
motivated researchers to explore how these metals interact dynamically with equity markets,
especially during financial crises and geopolitical shocks.
2.2.2.1
Hedging and Safe-Haven Properties
Banerjee and Pradhan (2024) examined how well precious metals protected investors
during the COVID-19 pandemic. Using a copula-based model, they found that gold remained
the most consistent safe-haven, offering strong protection during sharp equity selloffs. Silver
also showed some hedging potential, but its industrial use, especially in electronics and solar
panels, reduced its reliability during supply chain disruptions. Platinum and palladium, while
not traditional safe havens, provide value in specific sectors, such as the automotive and clean

17

energy industries, where they are used in catalytic converters and hydrogen fuel cells (Erdoğan
et al., 2022).
Lucey and Li (2015) also analyzed the behavior of precious metals over time. They
showed that gold typically retains its safe-haven status regardless of the business cycle, while
silver and platinum behave differently depending on macroeconomic conditions. For example,
silver’s correlation with stocks increases during economic booms, making it less reliable when
equity risk rises. This highlights the need for flexible portfolio strategies that account for the
changing behavior of each metal.
2.2.2.2
Volatility Spillovers and Asymmetric Effects
Precious metals do not just hedge risk, they also transmit volatility. Mensi et al. (2020)
found that metals like gold and silver sometimes send shocks into energy markets, particularly
during global events like the oil price crash 2020. Their study challenges the idea that precious
metals always reduce risk. Erdoğan et al. (2022) also found asymmetric volatility spillovers
between precious metals and clean energy stocks. Their results showed that safe-haven
behavior is stronger during severe downturns, especially for gold and silver, while metals like
palladium closely track equity market stress due to their reliance on industrial demand.
2.2.2.3
Macroeconomic Forces and Financialization
The behavior of precious metals is not only shaped by their inherent properties but also
by macroeconomic trends and financial market dynamics. Gold, for instance, often responds
strongly to inflation expectations and interest rate movements, especially during times of
market stress. Dinh et al. (2022) showed that long-term price movements of gold correlate

18

closely with macroeconomic factors like monetary policy and inflation, while short-term
fluctuations are driven by investor sentiment and geopolitical news.
Geopolitical events intensify this sensitivity. As (Baur & Smales, 2018) explain, gold
doesn’t just respond to actual conflicts, it tends to rise in value when investors fear something
might go wrong. A clear example is the Russia–Ukraine war in early 2022, when gold prices
jumped as markets looked for safer ground (World Bank, 2022). This shows how gold reflects
both global risk sentiment and financial instability.
Other precious metals, like palladium, react differently because of their industrial
importance and concentrated sources of supply. Russia, which produces close to 40% of the
world’s palladium, was hit with major sanctions after its invasion of Ukraine. This immediately
raised concerns about supply, and palladium prices soared to record levels in March 2022
(Hobson, 2022). It’s a strong reminder of how quickly markets can react when key resources
are at risk.
Financialization has recently changed how precious metals like gold and palladium
are traded in markets. They are no longer just physical commodities but have become key
financial market parts. Today, people invest in them through exchange-traded funds (ETFs),
futures, and other financial tools. As Tang and Xiong (2012) explain, this shift means their
prices are now more influenced by investor behavior and money flows than traditional supply
and demand.
During the COVID-19 pandemic, there was a significant increase in gold ETF
purchases, rising by over 25%. Investors sought safe places to put their money as the stock
market fell (Jégourel, 2021). At the same time, speculation around palladium futures increased

19

during the Russia–Ukraine conflict. This led to significant price swings and caused a gap
between its market price and the actual supply and demand for palladium (Ding et al., 2021).
Therefore, precious metals now serve two roles: gold remains a haven during economic
shocks, while palladium shows how financial speculation can create big market moves. This
adds new challenges for investors, who must now consider global economic trends and
financial market dynamics.
2.2.2.4
Precious vs. Strategic Metals: A Key Difference
Precious metals behave differently from strategic metals like lithium or cobalt. While
strategic metals tend to move in the same direction as clean energy and EV stocks (because
they are part of the same supply chain), precious metals, especially gold, often move in the
opposite direction during crises (Cagli, 2023). Silver, however, sits somewhere in between, as
it is both a precious and industrial metal. Its dual nature creates mixed patterns of volatility and
correlation that combine features of both groups (Y. Chen et al., 2022).
Therefore, literature underscores that precious metals continue to play a vital role in
financial markets. Gold continues to be a go-to asset for protection during crises, while other
metals offer additional benefits depending on industrial demand and market conditions.
However, financialization, volatility, and geopolitical shocks have made their behavior more
complex and less predictable. Understanding these changing dynamics is essential, especially
when comparing their role with that of strategic metals in modern portfolio construction.
2. 3 Hedging Effectiveness of Metals in Equity Portfolios
The ability of metals to serve as hedging instruments has been a significant topic in
financial economics, especially during times of heightened market uncertainty. While precious
metals such as gold and silver have traditionally been viewed as safe-haven assets, the evolving
role of strategic (non-precious) metals, like cobalt, nickel, and copper, has introduced new
20

dimensions to hedging strategies, particularly in the context of clean energy and EV-linked
portfolios.
Gold has long been recognized for its strong hedging properties. (Baur & Lucey, 2010a)
showed that gold can act as both a hedge and a safe-haven depending on market conditions,
especially during equity downturns. Similarly, Beckmann et al. (2019) found that gold provides
better protection when market volatility is high, making it a reliable asset for managing extreme
risk. However, the performance of other precious metals such as silver, platinum, and
palladium is less consistent, often influenced by their industrial demand (Erdoğan et al., 2022).
More recent studies have investigated the hedging potential of strategic metals.
Gustafsson et al. (2022) analyzed the role of cobalt and nickel in clean energy portfolios and
concluded that, under certain market conditions, these metals could partially hedge equity risk,
especially in the clean technology sector. Rajwani et al. (2023) further support this by showing
that portfolios combining energy metals with clean energy stocks tend to have lower downside
risk, provided that dynamic hedging strategies are used.
Dynamic correlation models such as the DCC-GARCH framework have been widely
adopted in evaluating hedge effectiveness. These models account for time-varying
relationships, offering a more realistic picture of how correlations evolve, especially during
periods of crisis. (Y. Chen et al., 2022) applied this approach to analyze hedge ratios between
metals and clean energy indices and found that strategic metals exhibited considerable hedging
power, though less stable than gold.
So, the evidence suggests that while precious metals remain the dominant hedge assets
due to their liquidity and historical performance, strategic metals have emerged as
complementary tools, particularly in portfolios that are heavily exposed to clean energy and

21

EV sectors. Their value in a hedging context may not lie in stability alone but in their ability
to track the economic pulse of the green transition.
2. 4 Methodologies in the Literature: DCC-GARCH, VAR, Granger Causality
Recent empirical research in financial econometrics has increasingly relied on dynamic
models to analyze how financial assets interact over time, especially during periods of
economic uncertainty. In the context of metals and equity markets, such as clean energy,
electric vehicle (EV), and traditional automakers, five key methodologies are widely used:
DCC-GARCH, Vector Autoregression (VAR), Granger Causality, Total Connectedness Index
(TCI), and Impulse Response Functions (IRF). This section reviews the application of these
models in the literature and outlines their main strengths and limitations.
2.4.1

Dynamic Conditional Correlation GARCH (DCC-GARCH)

The Dynamic Conditional Correlation GARCH (DCC-GARCH) model, developed by
(Engle, 2002a), has become a cornerstone in modeling time-varying correlations between
financial assets. Unlike static models, DCC-GARCH accounts for the evolving nature of
market relationships, making it particularly useful during crisis periods when asset correlations
often intensify. For instance, (Y. Chen et al., 2022) employed this model to examine how
spillovers from non-ferrous metals affect clean energy stock indices, while Gustafsson et al.
(2022) found that energy metals such as nickel and cobalt serve as dynamic hedging tools for
clean energy portfolios. More recently, Bastianin et al. )2025) documented how lithium and
cobalt correlations surged amid supply chain disruptions, and Cagli (2023) used the model to
reveal asymmetric volatility spillovers from battery metals to EV manufacturers.
The key strengths of DCC-GARCH lie in its ability to model time-varying correlations,
separate volatility from co-movement dynamics, and provide dynamic hedge ratios, which are

22

essential for managing financial risk in volatile markets. However, the model also has
limitations. When applied to large systems with many asset pairs, it can be computationally
intensive. The standard DCC specification also assumes symmetric responses to shocks, which
may not reflect real-world conditions unless extended (e.g., through asymmetric DCC or
regime-switching variants). Finally, DCC-GARCH models co-movement but does not infer
directionality or causality, requiring complementary models such as VAR or Granger causality
for full interpretability.
2.4.2

Vector Autoregression (VAR)

The Vector Autoregression (VAR) model, initially introduced by SIMS (1980), has
become one of the most widely used techniques in financial and macroeconomic research for
studying dynamic relationships among multiple time series. Unlike traditional single-equation
models, VAR treats every variable in the system as endogenous, allowing for a flexible
structure where each variable is explained by its lags and the lags of all other variables in the
system. This makes VAR particularly suitable for analyzing interconnected markets, where
shocks in one market (e.g., strategic metals) can ripple across others, such as electric vehicle
(EV) stocks or clean energy indices.
Early studies applying VAR to financial markets include Campbell and Shiller (1988),
who used VAR to test the present-value model of stock prices by linking stock prices,
dividends, and earnings. Schwert (1989) employed VAR to explore why stock market volatility
changes over time, highlighting the role of macroeconomic factors. Similarly, Hasbrouck
(1991)) utilized VAR to measure the informational content of stock trades, providing insights

23

into the microstructure of financial markets. These foundational works established VAR as a
cornerstone for empirical analysis in finance.
More recent research has extended the application of VAR models to analyze the
dynamic linkages between commodity markets and financial assets, particularly in strategic
metals and clean energy equities. Alekseev et al. (2024) applied a VAR framework to
investigate return and volatility spillovers between battery metals (such as lithium and cobalt)
and EV manufacturers, finding significant transmission effects from upstream material markets
to downstream stock prices. Similarly, Mensi et al. (2022) used a VAR-based connectedness
approach to analyze volatility spillovers between strategic commodities such as lithium and
US-based sectoral stock indices, emphasizing the evolving influence of critical minerals on
equity markets. Furthermore, (Y. Chen et al., 2022) employed VAR models to explore the
hedging effectiveness of non-ferrous metals against clean energy stocks, demonstrating that
metals such as copper and nickel can act as short-term hedges during market stress periods.
The major strengths of VAR models lie in their ability to capture mutual feedback
effects, their flexible structure without requiring strong theoretical restrictions, and their
usefulness in forecasting and policy analysis. However, VAR models also have limitations.
They typically require that the underlying time series be stationary, and in the presence of many
variables, they risk overfitting unless model selection criteria are carefully applied.
Additionally, while VAR can detect temporal relationships, it does not inherently establish
causal directions unless complemented by tools like Granger causality tests or impulse
response functions.

24

2.4.3

Granger Causality

Granger causality, introduced by Granger ) 1969) is a widely used statistical concept
in time-series econometrics that helps determine whether one time series can predict another.
Unlike traditional notions of causality, Granger causality is based on predictability: if past
values of variable X contain information that helps forecast future values of variable Y, beyond
the information contained in past values of Y itself, then X is said to “Granger-cause” Y. This
framework has become fundamental in finance and macroeconomics for investigating
directional linkages between markets, assets, and macroeconomic indicators.
For instance, (N.-F. Chen et al., 1986) applied Granger causality tests to assess whether
economic forces like industrial production and inflation Granger-cause stock market returns,
finding significant predictive relationships. Similarly, Bekaert and Harvey (1995) examined
causality patterns between developed and emerging stock markets, revealing how information
flows across global markets during periods of volatility. More recently, Billio et al. (2012)
employed Granger causality networks to study systemic risk and interconnectedness among
hedge funds, banks, and insurance companies, highlighting the growing complexity of
financial linkages.
In recent years, Granger causality has provided valuable insights into how shocks
originating from metals such as gold, silver, or cobalt can influence equity sectors like clean
energy or electric vehicles in commodity and equity markets. Erdoğan et al. (2022)used
Granger causality tests to analyze the safe-haven properties of precious metals relative to clean
energy equities, finding asymmetric predictive relationships under different market regimes.
Similarly, Ghosh et al. (2023) applied quantile-based Granger causality techniques to study

25

tail-risk dependencies between metals and energy markets during the COVID-19 pandemic,
revealing heightened contagion during crises.
The main strengths of Granger causality analysis include its simplicity, interpretability,
and ability to detect directional predictability in time-series relationships. However, it also has
limitations. It does not imply true economic causality, being purely statistical; its results can
be sensitive to lag length selection; and it assumes linear relationships unless extended to
nonparametric or quantile-based frameworks.
One of the main strengths of the TCI approach is its ability to turn the complexity of
financial systems into a clear and understandable measure. It also allows researchers to monitor
how market interconnectedness evolves over time and to identify which markets are major
transmitters or receivers of volatility. However, the method is not without limitations. TCI
results heavily depend on how well the underlying VAR model is specified, and they can be
sensitive to choices like lag length and sample size. Moreover, while TCI can show how much
markets are connected, it does not by itself reveal the direction of influence; additional
techniques like Granger causality tests or impulse response analysis are needed to uncover
those relationships.
2.4.4

Total Connectedness Index (TCI)

The Total Connectedness Index (TCI), introduced by (Diebold & Yilmaz, 2012),
quantifies systemic risk spillovers and interconnectedness among financial assets. Unlike static
correlation measures, TCI dynamically tracks how shocks in one asset explain forecast error
variances in others over time, offering critical insights into market contagion during crises.
The TCI methodology originates from variance decomposition in a Vector
Autoregression (VAR) framework. By breaking down the forecast error variance of each

26

variable into parts attributable to shocks from itself and shocks from others, researchers can
measure both own-market dependence and cross-market spillovers. The aggregation of crossvariable spillovers leads to calculating the total connectedness index, offering a clear singlenumber summary of overall market interdependence.
Early influential applications include (Diebold & Yilmaz, 2012), who used TCI to
examine global equity market interconnectedness across developed and emerging economies,
demonstrating that spillovers intensified during the 2008 financial crisis. Baruník and&
Křehlík (2018) extended the methodology into the frequency domain, allowing researchers to
decompose connectedness across different time horizons, such as short-term and long-term
spillovers. These innovations have made TCI a core tool in analyzing financial contagion,
portfolio risk, and systemic instability.
Recent studies have increasingly applied TCI to commodity markets, especially in the
context of strategic metals and clean energy stocks. For instance, Ghosh et al. (2023) used
quantile-based TCI to analyze how energy and metal markets became more interconnected
during the COVID-19 pandemic, finding that the intensity of connectedness varies across
quantiles of return distributions. Similarly, Cagli (2023) applied frequency-based TCI to trace
how battery metals like lithium and cobalt transmit volatility to EV manufacturers and clean
energy equities, highlighting that spillovers are stronger during market stress.
2.4.5

Impulse Response Functions (IRF)
Impulse Response Functions (IRFs) are critical tools in time-series econometrics for

analyzing how shocks to one variable dynamically propagate through a system of interrelated
variables. Derived from Vector Autoregressive (VAR) models, IRFs trace the time path of an
endogenous variable’s response to an exogenous shock, isolating its magnitude, direction, and

27

persistence (C. A. Sims, 1980). This method is particularly valuable for studying the
interdependencies between strategic metals (e.g., lithium, cobalt), clean energy equities, and
traditional automaker stocks, especially during episodes of heightened volatility or geopolitical
shocks. For instance, IRFs allow researchers to answer practical questions such as: How does
a sudden increase in cobalt prices affect Tesla’s stock returns over the following ten days? What
is the dynamic path of clean energy sector volatility following a shock in gold prices?
IRFs operate under the assumption that shocks are unexpected innovations that hit one
variable at a specific point in time and ripple through the system across subsequent periods.
Typically, to properly isolate these causal pathways, shocks are orthogonalized using Cholesky
decomposition, which ensures that residuals (errors) are uncorrelated (Lütkepohl, 2005).
However, one important limitation of the traditional Cholesky-based IRFs is that results can
be highly sensitive to the ordering of variables in the VAR system, which can bias
interpretations. To overcome this issue, Generalized Impulse Response Functions (GIRFs),
proposed by Pesaran and Shin )1998) , offer an alternative by generating shock responses that
are invariant to variable ordering.
In empirical research on commodity and equity markets, IRFs have been extensively
used to capture shock transmission mechanisms. For example, Pakrooh and Manera (2024)
utilized impulse response analysis to examine how shocks in strategic metal prices affect stock
and carbon markets in the European Union. Alekseev et al. (2024) also investigated return and
volatility spillovers between battery metals and EV manufacturer stocks, emphasizing the need
to understand how raw material price movements dynamically influence downstream equity
valuations.

28

Despite their versatility, IRFs have notable limitations. Besides sensitivity to
identification schemes, standard IRFs often assume linearity and symmetry in shock responses,
which may not hold during financial crises or structural breaks. Advanced versions such as
Quantile IRFs or Local Projection IRFs have been proposed in recent literature to address these
concerns.
2. 5 Gaps in the Literature and Research Justification
Despite the growing body of research exploring the linkages between metals and
financial markets, several important gaps remain unaddressed. First, in isolation, most existing
studies examine either precious metals (such as gold and silver) or strategic metals (such as
lithium, cobalt, and nickel). Few papers offer a comparative analysis across multiple equity
sectors, including electric vehicle (EV) manufacturers, traditional automakers, and clean
energy indices, which limits the ability to understand how different classes of metals interact
with distinct parts of the financial market.
Second, although dynamic models like DCC-GARCH and VAR have been increasingly
adopted, a significant portion of the literature still relies on static models or short time windows
that fail to capture how asset interlinkages evolve during major systemic shocks, such as the
COVID-19 pandemic and the Russia–Ukraine war. As Ghosh et al. (2023) and Cagli (2023)
highlight, market connectedness, and volatility spillovers can shift dramatically during periods
of global stress, emphasizing the need for more crisis-sensitive, dynamic analyses.
Third, while some research has examined hedging properties of precious metals
against clean energy equities like Erdoğan et al. (2022), there is a notable lack of
comprehensive hedging effectiveness comparisons between strategic and precious metals,
especially regarding their performance across clean energy, EV, and traditional automotive
29

sectors. Most studies evaluate hedging behavior either at the market-wide level or for a limited
subset of assets, without investigating sectoral asymmetries.
Fourth, although Granger causality tests, Impulse Response Functions (IRFs), and
connectedness measures (TCI) are individually well-established, few studies combine these
techniques into a comprehensive, integrated framework. A dynamic, multi-method approach is
crucial for uncovering how shocks originate, propagate, and impact asset classes differently
over time.
Finally, while global crises have undoubtedly altered commodity and financial market
behavior, relatively little work explicitly incorporates events like the COVID-19 pandemic and
geopolitical disruptions into the modeling of metals–equity relationships. However, as the
International Energy Agency IEA (2021) and the World Bank (2022) noted, supply chain
vulnerabilities and strategic metal dependencies have become increasingly critical in the postpandemic global economy.
Given these gaps, this thesis justifies its contribution by proposing a comprehensive,
dynamic, and differentiated analysis at the sectoral level of strategic and precious metals in
relation to clean energy equities, electric vehicle (EV) manufacturers, and traditional
automaker stocks. By employing a combination of DCC-GARCH, VAR, Granger causality,
TCI, and IRF methodologies over a sample period that includes major global crises, this study
offers fresh empirical insights into hedging effectiveness, market interconnectedness, and risk
transmission channels in an era of accelerated energy transition and geopolitical uncertainty.
In summary, this chapter provided an overview of the financialization of strategic and
precious metals, their evolving linkages with clean energy equities, electric vehicle (EV)
30

manufacturers, and traditional automakers, and the methodologies commonly used to analyze
these dynamics. It highlighted key empirical findings and revealed several research gaps that
motivate this study’s comprehensive and dynamic approach.
Chapter Three: RESEARCH METHODOLOGY
3.1 Research Design
This study employs a quantitative, time-series-based research design to explore how
strategic and precious metals interact with equity markets, particularly clean energy indices,
electric vehicle (EV) manufacturers, and traditional automakers. Given the non-stationary and
crisis-sensitive nature of financial relationships, especially during periods of uncertainty such
as the COVID-19 pandemic and the Russia–Ukraine war, a dynamic modeling framework is
employed to capture time-varying correlations, volatility spillovers, causality patterns, and risk
transmission channels.
The empirical methodology integrates several advanced econometric techniques,
including Dynamic Conditional Correlation GARCH (DCC-GARCH) models, Vector
Autoregression (VAR), Granger causality tests, Total Connectedness Index (TCI), and Impulse
Response Functions (IRF). By employing a multi-method approach, the study ensures a
comprehensive understanding of the evolving relationships between metals and equity markets
under normal and crisis conditions.
The analysis is conducted using daily return data spanning from January 2014 to
January 2024. This period captures multiple phases of financial market behavior, including the
pre-pandemic era, the COVID-19-induced market volatility, and the geopolitical disruptions
arising from the Russia–Ukraine war. Such a rich dataset allows for a robust examination of

31

dynamic risk transmission mechanisms and offers practical insights for portfolio
diversification and hedging strategies in sustainability-driven investments.
The chapter proceeds by detailing the data collection process, variable selection
criteria, model specifications, estimation procedures, diagnostic tests, and the rationale behind
the methodological choices employed in this study.
3.2 Data Collection and Variable Selection
This study employs a comprehensive dataset spanning January 31, 2014, to January 31,
2024, to analyze the dynamic interactions between strategic metals, precious metals, and equity
markets. The dataset captures both normal market conditions and systemic shocks, including
the COVID-19 pandemic and the Russia–Ukraine conflict, ensuring a robust examination of
volatility spillovers and hedging effectiveness.
3.2.1

Data Sources and Variable Definitions

Data was sourced from S&P Capital IQ, Bloomberg, London Metal Exchange (LME),
Westmetall, and currency exchange platforms to ensure accuracy and consistency. The
variables are categorized as follows:
1. Equity Markets
•

•

EV Manufacturers:
▪

Tesla, Inc. (NasdaqGS: TSLA)

▪

BYD Company Limited (SEHK: 1211)

Traditional Automakers:
▪

Honda Motor Co., Ltd. (TSE: 7267)

▪

Toyota Motor Corporation (TSE: 7203)

▪

General Motors Company (NYSE: GM)

▪

Ford Motor Company (NYSE: F)
32

•

Clean Energy Indices:
▪

NASDAQ Clean Edge Green Energy Index (^CELS)

▪

S&P Global Clean Energy Index (^SPGTCLEN)

▪

WilderHill Clean Energy Index (ECO)

2. Precious Metals (USD per troy ounce):
•

Gold (Day Close Price)

•

Silver (Day Close Price)

•

Palladium (Day Close Price)

•

Platinum (Day Close Price)

3. Strategic Metals (USD per metric ton):
•

Copper (LME Cash Settlement)

•

Aluminium (LME Cash Settlement)

•

Nickel (LME Cash Settlement)

•

Zinc (LME Cash Settlement)

•

Cobalt (LME: ^CO Spot Close Price)

4. Currency Exchange Rates:
•

CAD/USD, CNY/USD, JPY/USD, HKD/USD (for standardizing international
stock prices into USD).

3.2.2

Data Processing
3.2.2.1 Currency Conversion:
▪

Foreign-denominated stock prices (e.g., BYD in HKD, Honda/Toyota in JPY)
were converted to USD using daily exchange rates to ensure comparability.

33

▪

USD

Example: BYD’s HKD price converted PUSD = PHKD × HKD rate

3.2.2.2 Handling Missing Data:
▪

Gaps in metal prices (e.g., cobalt) and equity series were addressed via forwardfilling, assuming the last observed value persisted until new data became
available.

3.2.2.3 Return Calculation:
▪

Daily log returns computed as:
(1)

P

rt =ln (P t ) × 100
t−1

This transformation stabilizes variance and mitigates non-stationarity.
3.2.2.4 Alignment:
▪

All

series

aligned

to trading

days (NYSE

calendar)

to

exclude

weekends/holidays and ensure temporal consistency.
3.2.3

Rationale for Variable Selection

The selection of variables in this study was guided by both theoretical relevance and
practical importance.
Strategic metals such as Copper, Aluminium, Nickel, Zinc, and Cobalt were selected
due to their essential role in clean technologies, especially in electric vehicle batteries,
renewable energy infrastructure, and energy storage systems (IEA, 2021). These metals are
increasingly recognized as pivotal indicators of the global energy transition.
Precious metals including Gold, Silver, Platinum, and Palladium, were incorporated
given their longstanding reputation as safe-haven assets and their hedging potential during
34

market turbulence (Baur & Lucey, 2010b; Erdoğan et al., 2022). Their inclusion allows for a
comparative analysis between strategic and traditional safe-haven metals across different
equity sectors.
Clean energy indices like CELS, SPGTCLEN, and ECO, were chosen to represent
diversified investment portfolios tied to renewable energy, energy storage, and sustainable
technologies. They provide a broader sectoral view beyond individual firm-level dynamics and
are vital for studying systemic spillovers and hedging performance within the green economy.
EV manufacturers (Tesla, BYD) and traditional automakers (Toyota, Honda, GM,
Ford) were included to allow a sectoral comparison between firms with high versus moderate
exposure to critical minerals. This differentiation enables a deeper understanding of how
metals-linked volatility transmits differently across industries.
Exchange rates were incorporated to accurately convert non-USD denominated stock
prices into a consistent currency base. This step was crucial for ensuring comparability across
international assets and avoiding distortions caused by foreign exchange fluctuations.
3.3 Model Specifications
This section details the econometric models employed to analyze the dynamic
relationships between strategic metals, precious metals, and equity markets. The specifications
are designed to address research questions on time-varying correlations, volatility spillovers,
causal linkages, systemic connectedness, and shock propagation.
3.3.1

Dynamic Conditional Correlation GARCH (DCC-GARCH)

The Dynamic Conditional Correlation Generalized Autoregressive Conditional
Heteroscedasticity (DCC-GARCH) model is a powerful econometric tool designed to analyze
time-varying correlations and volatilities among multiple financial time series. It is particularly
effective in capturing how correlations evolve over time. It helps in providing evidence into
35

dynamic relationships within volatile markets, such as the interplay between precious and nonprecious metals, automaker stock volatilities, and clean energy trends. The model is
implemented in two stages: first, by estimating individual variances using univariate GARCH
models, and then by calculating dynamic correlations between variables (Engle, 2002b;
Sadorsky, 2012).
The DCC-GARCH model can be divided into two key components:
1. Conditional Mean Equation (ARMA (p, q)):
This models the return of each time series using an Autoregressive Moving Average (ARMA)
process:
p

q

(2)

rt = µt + ∑ ᶲi rt−i + ∑ θj εt−j + εt
i=1

j=1

Where:
•

rt : Asset return at time t.

•

µt : Conditional means of the return series.

•

ᶲi : Autoregressive (AR) coefficients.

•

θj : Moving average (MA) coefficients.

•

εt : Residual error term.

2. Conditional Variance and Correlation Estimation:
This section consists of two steps:
Step 1: Variance Estimation (GARCH (1,1))
The conditional variance of each return series is estimated using the GARCH (1,1)
model:
ht = ω + αε2t−1 + βht−1
Where:
•

ht : Conditional variance at time t.

•

ω: Constant term (baseline volatility).

•

α: Short-term volatility persistence (ARCH effect).
36

(3)

•

β: Long-term volatility persistence (GARCH effect).

•

Constraints: ω >0, α, β ≥ 0, α + β <1.

Step 2: Dynamic Conditional Correlations (DCC):
The time-varying correlations between variables are estimated as:
𝑄𝑡 = (1 − 𝑎 − 𝑏)𝑄̅ + 𝑎(𝑧𝑡−1 𝑧́𝑡−1 ) + 𝑏𝑄𝑡−1

(4)

Where:
•

𝑄𝑡 : Time-varying covariance matrix of standardized residuals.

•

𝑄̅ : Unconditional covariance matrix of 𝑧𝑡 .

•

𝑎 : Short-term effect of past shocks.

•

𝑏 : Long-term effect of past correlations.
ε
zt : Standardized residuals (zt = t⁄ )
√ht

•

The final dynamic conditional correlation matrix is calculated as:
−1

−1

R t = diag(Qt ) 2 Qt diag(Qt ) 2

(5)

The final dynamic conditional covariance matrix is calculated as:
(6)

Ht = Dt R t Dt
Where 𝐷𝑡 = diag (√ℎ1,𝑡 , √ℎ2,𝑡 , … , √ℎ𝑛,𝑡 ) contains GARCH volatilities.

This study employs the DCC-GARCH model to analyze the dynamic relationships
between multiple financial assets. Specifically, it investigates (1) the evolving correlations
between strategic metals (Copper, Aluminium, Nickel, Zinc, Cobalt) and clean energy indices
(ECO, CELS, SPGTCLEN), (2) the interactions between precious metals (Gold, Silver,
Platinum, Palladium) and electric vehicle (EV) manufacturers (Tesla, BYD), and (3) the
sensitivity of traditional automakers (Toyota, Honda, GM, Ford) to fluctuations in metal prices.
The objective is to differentiate the financial behavior of various asset classes under
different market conditions, including periods of systemic stress such as the COVID-19

37

pandemic and the Russia–Ukraine war. This approach helps identify which metals serve as
effective hedging tools, how volatility is transmitted across markets, and whether clean energy
stocks and EV manufacturers are more sensitive to metal price fluctuations than traditional
automakers.
The results from the DCC-GARCH model serve as a foundation for subsequent
analyses, including Vector Autoregression (VAR) in Section 3.3.2. The VAR model will further
explore the causal relationships between these assets, providing a comprehensive
understanding of how shocks propagate across strategic metals, precious metals, clean energy
indices, and automotive stocks. This integrated approach ensures a more detailed analysis of
financial interconnectedness across sectors.
3.3.2

Vector Autoregression (VAR)

Vector Autoregression (VAR) is an econometric model that captures linear
interdependencies among multiple time series variables (Lütkepohl, 2005; C. A. Sims, 1980) .
In this study, the VAR model is employed to analyze the dynamic interactions between strategic
metals (Copper, Aluminium, Nickel, Zinc, Cobalt), precious metals (Gold, Silver, Platinum,
Palladium), and equity markets, including clean energy indices (ECO, CELS, SPGTCLEN),
electric vehicle (EV) manufacturers (Tesla, BYD), and traditional automakers (Toyota, Honda,
GM, Ford).
A standard VAR model of order p is specified as follows:
𝑦𝑡 = 𝜈 + 𝐴1 𝑦𝑡−1 + 𝐴2 𝑦𝑡−2 + ⋯ + 𝐴𝑝 𝑦𝑝−1 + 𝑢𝑡

(7)

Where:
•

𝑦𝑡 is a vector of endogenous variables (e.g., metal prices and stock returns).
𝑦𝑡 = (𝑦1𝑡 , … , 𝜈𝑘𝑡 )′ ,

𝑦𝑡 =[

𝑚𝑒𝑡𝑎𝑙𝑡
]
𝑠𝑡𝑜𝑐𝑘 𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑡
38

(8)

•

𝜈 𝑖𝑠 𝑎 𝑣ector of intercept terms. 𝜈 = (𝜈1 , … , 𝜈1 )′

•

𝐴1 , 𝐴2 , … , 𝐴𝑝 are coefficient matrices for lagged terms.

•

𝑢𝑡 is a white noise error vector with E (𝑢𝑡 )=0 and E (𝑢𝑡 𝑢́ 𝑡 )=∑ 𝑢, and E
(𝑢𝑡 𝑢́ 𝑠 )=0, s ≠ t
ut = (u1t , … , ukt )′

(9)

The optimal lag length p for the VAR model is determined using the Akaike Information
Criterion (AIC), which is defined as:
(10)

𝐴𝐼𝐶 = 2𝑘 − 2𝑙𝑛 (𝐿̂)
Where:
•

𝑘 is the number of parameters, and 𝐿̂ is the maximized likelihood of the model.

•

In this study, the AIC is calculated using the VARselect() function in R, which
automatically identifies the optimal lag length for each VAR model.

The VAR model is considered stable if all eigenvalues of the characteristic polynomial
lie outside the unit circle:
𝑑𝑒𝑡(𝐼𝑘 − 𝐴1 𝑧 − ⋯ − 𝐴𝑝 𝑧 𝑝 ) ≠ 0

for (∣z∣≤1).

(11)

The VAR model is chosen because it captures dynamic interactions between multiple
time series without assuming causality (Lütkepohl, 2005; C. A. Sims, 1980). It provides a
flexible framework for analyzing how shocks in one market (e.g., strategic metals) affect others
(e.g., EV manufacturers, clean energy indices). Unlike single-equation models, VAR treats all
variables as endogenous, enabling a comprehensive view of feedback effects between asset
classes. This is essential for understanding the bidirectional relationships between metal prices

39

and stock returns, especially during crises like the COVID-19 pandemic and the Russia–
Ukraine conflict.
3.3.3

Granger Causality Tests

Granger causality tests, introduced by Granger (1969), determine whether one time
series can predict another. The idea is that a cause cannot come after the effect. Thus, if a
variable X affects a variable Y, the former should help improve the predictions of the latter
variable. In this study, Granger causality is tested to determine whether metal prices Grangercause stock returns of EV manufacturers and traditional automakers (or vice versa).
The Granger Causality test is based on the following regression model:
1. Unrestricted Model (EV returns regressed on lagged metals and own lags):
To test whether metal prices Granger-cause stock returns (or vice versa), we estimate:
𝑌𝑡 =𝛼 + ∑𝑝𝑖=1 𝛽𝑖 𝑌𝑡−𝑖 + ∑𝑝𝑖=1 𝛾𝑖 𝑋𝑡−𝑖 +𝜀𝑡

(12)

𝑠𝑡𝑜𝑐𝑘 𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑡 =𝛼 + ∑𝑝𝑖=1 𝛽𝑖 𝑠𝑡𝑜𝑐𝑘 𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑡−𝑖 + ∑𝑝𝑖=1 𝛾𝑖 𝑚𝑒𝑡𝑎𝑙𝑡−𝑖 +𝜀𝑡

(13)

2. Restricted Model (excluding metal lags):
𝑌𝑡 =𝛼 + ∑𝑝𝑖=1 𝛽𝑖 𝑌𝑡−𝑖 + 𝜀 ′ 𝑡

(14)

𝑠𝑡𝑜𝑐𝑘 𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑡 =𝛼 + ∑𝑝𝑖=1 𝛽𝑖 𝑠𝑡𝑜𝑐𝑘 𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑡−𝑖 +𝜀 ′ 𝑡

(15)

The null hypothesis is that 𝑃𝑎𝑠𝑡 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑋𝑡 does not Granger-cause 𝑌𝑡
𝐻0 : 𝛾1 = 𝛾2 = ⋯ = 𝛾𝑝 = 0 (No Granger causality)

(16)

Therefore, Sum squared Residual from model are calculated as follows:
𝑅𝑆𝑆0=∑𝑇𝑡=1 𝜀̂𝑡2

(17)

2

(18)

𝑅𝑆𝑆1=∑𝑇𝑡=1 𝜀̂′ 𝑡

𝑅𝑆𝑆0 is the residual sum of squares from the restricted model (Equation 13), while
𝑅𝑆𝑆1 is from the unrestricted model (Equation 15).
The F-test compares the restricted (no causality) and unrestricted models:

40

(𝑅𝑆𝑆0 − 𝑅𝑆𝑆1 )⁄
𝑃
𝐹=
𝑅𝑆𝑆1
⁄(𝑇 − 2𝑝 − 1)
Ho is rejected if F > F0.05,(p,T−2p−1)

(19)

where:
•

𝑅𝑆𝑆0 and 𝑅𝑆𝑆1 are residual sums of squares from the restricted and unrestricted
models, respectively.

•
3.3.4

T is the sample size.
Impulse Response Functions (IRF)

Impulse Response Functions (IRF) are a key econometric tool used to trace the dynamic
response of one variable to a shock in another within a Vector Autoregression (VAR) model
(Lütkepohl, 2005). In this study, IRFs are used to analyze how shocks in strategic and precious
metals affect the stock returns of electric vehicle (EV) manufacturers, traditional automakers,
and clean energy indices.
The IRF is derived from the Vector Moving Average (VMA) representation of the VAR
model (Lütkepohl, 2005):
𝑦𝑡 =µ+ ∑∞
𝑠=0 𝛹𝑠 𝑢𝑡−𝑠

(20)

•

yt is a vector of endogenous variables in the Vector Autoregression (VAR) model.

•

µ is a vector of intercept terms (mean values) in the VAR model.

•

ut−s is a vector of white noise error terms.

•

Ψs is the s-th IRF matrix, capturing the impact of shocks.
𝛹𝑠 = 𝐽𝐹 𝑠 𝐽′

Where:
•
•

(21)

1 0
) is an identity matrix, which standardizes the shocks.
0 1
A ⋯ Ap
F= ( 1
) is a block matrix constructed from the coefficient matrices of
I
0
J=(

the VAR model.
41

•

A1 , . . . , Ap are the coefficient matrices of the VAR model.

The IRF for a shock in a variable j on variable i at horizon s is defined as:
𝛿𝑦

𝐼𝑅𝐹𝑖,𝑗 (𝑠) = 𝛹𝑠,𝑖𝑗 = 𝛿 𝑖,𝑡+𝑠

(22)

𝑢𝑗,𝑡

•

Ψs,ij measures the response of variable i (e.g., EV returns) at time t + s to a one-unit
shock in variable j (e.g., metal prices) at time t.
In this study, the IRFs are calculated using Cholesky decomposition, ensuring that the

shocks are orthogonalized (Lütkepohl, 2005). This means that each shock in the system is
treated as independent, preventing the overlap of effects from multiple shocks. Cholesky
decomposition is a method that transforms the residual covariance matrix of the VAR model
into a lower triangular matrix, making the shocks uncorrelated by construction. This approach
allows for clear interpretation of the impact of a shock in one variable (e.g., metal prices) on
another (e.g., EV stock returns), without interference from simultaneous shocks in other
variables (Hamilton, 1994).
3.3.5

Diebold and Yilmaz (2012) Approach

This study applies the volatility spillover framework developed by (Diebold & Yilmaz,
2012) to analyze the interconnectedness and transmission of volatility among strategic metals,
precious metals, electric vehicle (EV) manufacturers, traditional automakers, and clean energy
indices. This approach is based on the Generalized Forecast Error Variance Decomposition
(GFEVD) of a Vector Autoregressive (VAR) model, enabling a robust analysis of how volatility
propagates across sectors.
The analysis begins with a stationary VAR(M) model:
(23)

𝑌

𝑀𝑡 = ∑ 𝜌𝑖 𝑀𝑡−1 + 𝜀𝑡
Where:

𝑖=1

42

•

Mt : Vector of endogenous variables (e.g., oil prices, automaker stocks, clean energy
indices).

•

ρi : Coefficient matrices capturing lagged relationships.

•

εt : Vector of disturbance terms.
The model is transformed into its moving average (∞) representation to facilitate the

analysis of dynamic spillovers. To ensure order invariance, the Generalized Forecast Error
Variance Decomposition (GFEVD) method Koop et al. (1996) and Pesaran and Shin (1998) is
employed. This allows an unbiased decomposition of forecast error variance, with
contributions from each variable normalized to sum to one. The H‐step ahead GFEVD is
calculated as follows:

𝑖
𝜈𝑚𝑛
=

𝐻−1
µ−1
́ 𝑍ℎ ∑ 𝜋𝑛 )2
𝑛𝑛 ∑ℎ=0 (𝜋𝑚
∑𝐻−1
́ 𝑍ℎ ∑ 𝑍́ℎ 𝜋𝑚
ℎ=0 𝜋𝑚

(24)

The standard deviation of the error term for the nth equation is denoted as 𝜇𝑛𝑛 , and Ʃ
represents the variance matrix of the error vector. The selection vector, 𝜋𝑚 , takes the value of
1 for the πth element and 0 for all other elements. However, the sum of the replaced elements
in each row of the variance decomposition table may not add up to 1. As a result, each
component of the variance decomposition matrix is normalized as
(25)

(𝑖)

𝜈𝑚𝑛 (𝐻)

(𝑖)
𝜈̃𝑚𝑛 =

𝑁

∑
𝑛=1

Where ∑

N
n=1

(i)

(𝑖)

𝜈𝑚𝑛 (𝐻)

ν̃mn (H) = 1 and ∑

N

(i)

m,n=1

ν̃mn (H) = H.

43

The total volatility spillover index can be calculated as:
𝑁

𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟(𝐻) =

=

(26)

(𝑖)

∑

𝜈̃𝑚,𝑛=1 (𝐻)

𝑚,𝑛=1(𝑚≠𝑛)
⋅ 100
𝑁
(𝑖)
∑
𝜈̃𝑚,𝑛=1 (𝐻)
𝑚,𝑛=1
𝑁
(𝑖)
∑
𝜈̃𝑚,𝑛=1 (𝐻)
𝑚,𝑛=1(𝑚≠𝑛)

⋅ 100
𝑁
This quantifies the overall proportion of forecast error variance attributable to crossvariable spillovers. In our context, this reflects the degree to which oil price shocks influence
volatility in automotive and clean energy stocks and vice versa.
To assess directional spillovers, the indices for spillovers received by market i and
transmitted from market i are given by:
𝑁

𝜈̃𝑚𝑛 (𝐻)

𝑚,𝑛=1(𝑚≠𝑛)
𝑁
(𝑖)
∑
𝜈̃𝑚𝑛 (𝐻)
𝑚,𝑛=1
𝑁
(𝑖)
∑
𝜈̃𝑚𝑛 (𝐻)
𝑚,𝑛=1(𝑚≠𝑛)

𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟𝜋.𝑖 (𝐻) =

=

(27)

(𝑖)

∑

𝑁

⋅ 100

⋅ 100

Next, to identify the volatility spillovers that are transmitted from market i to other
markets, we use:
𝑁

𝜈̃𝑚𝑛 (𝐻)

𝑚,𝑛=1(𝑚≠𝑛)
𝑁
(𝑖)
∑
𝜈̃𝑚𝑛 (𝐻)
𝑚,𝑛=1
𝑁
(𝑖)
∑
𝜈̃𝑚𝑛 (𝐻)
𝑚,𝑛=1(𝑚≠𝑛)

𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟.𝜋𝑖 (𝐻) =

=

(28)

(𝑖)

∑

𝑁

44

⋅ 100

⋅ 100

To analyze the spillover of the overall volatility from a particular market to other
markets, we compute the difference between the gross volatility shocks transmitted to and
received from all other markets in the sample (Equation (29)).
𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟 𝑖 (𝐻) = 𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟𝜋.𝑖 (𝐻) − 𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟.𝜋𝑖 (𝐻)

(29)

Furthermore, it is also possible to determine the net pairwise volatility spillover
between two specific markets (e.g., m and n) by employing the approach detailed in Equation
(30):
𝑖 (𝐻)
𝑆𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟𝑚𝑛
=[

(𝑖)

̃𝑛𝑚 (𝐻)
𝜈
𝑁

∑

𝑚,𝑝=1

=[

(𝑖)

̃𝑚𝑝 (𝐻)
𝜈

(30)

(𝑖)

−

̃𝑚𝑛 (𝐻)
𝜈
𝑁

∑

𝑛,𝑝=1

(𝑖)

] ⋅ 100

̃𝑛𝑝 (𝐻)
𝜈

(𝑖)
(𝑖)
𝜈̃𝑛𝑚 (𝐻) − 𝜈̃𝑚𝑛 (𝐻)

] . 100
𝑁
In this study, the (Diebold & Yilmaz, 2012) volatility spillover framework is applied
using a rolling window of 250 trading days (approximately one year) and a forecast horizon of
200 days (HHH-step = 200). This configuration enables the capture of the dynamic nature of
volatility transmission across strategic metals (Copper, Aluminium, Nickel, Zinc, Cobalt),
precious metals (Gold, Silver, Platinum, Palladium), clean energy indices (ECO, CELS,
SPGTCLEN), electric vehicle (EV) manufacturers (Tesla, BYD), and traditional automakers
(Toyota, Honda, GM, Ford). The optimal lag length for the underlying Vector Autoregression
(VAR) model is determined using the Akaike Information Criterion (AIC), ensuring that the
model is both parsimonious and statistically robust. Generalized Forecast Error Variance
Decomposition (GFEVD) is employed to achieve order-invariant spillover estimates,
enhancing the reliability of the results. The analysis is conducted on daily data spanning from
January 31, 2014, to January 31, 2024, covering both normal and crisis periods, including the
COVID-19 pandemic and the Russia–Ukraine conflict. Volatility for each asset is calculated

45

using the absolute value of daily log returns, providing a consistent measure of risk
transmission across the selected markets.
3.4 Limitations and justifications of Selected Methodologies
While this study employs advanced econometric techniques, including DCC-GARCH,
Vector Autoregression (VAR), Granger Causality Tests, Impulse Response Functions (IRF),
and the (Diebold & Yilmaz, 2012) Volatility Spillover Index, several limitations should be
considered. First, the DCC-GARCH model, despite its ability to capture time-varying
correlations, assumes symmetric volatility responses to shocks, which may not fully reflect
market behavior during extreme events (Engle, 2002b). Second, the VAR model, although
powerful for capturing dynamic relationships, is sensitive to the choice of lag length and
requires the data to be stationary, potentially leading to a loss of long-term information when
differencing is applied (Lütkepohl, 2005). Third, Granger causality tests, while useful for
detecting predictive relationships, do not establish true causal connections (Granger, 1969),
and the Impulse Response Functions (IRFs) rely on Cholesky decomposition, making them
sensitive to variable ordering (Lütkepohl, 2005). Finally, the (Diebold & Yilmaz, 2012)
framework is influenced by the choice of rolling window (250 days) and forecast horizon (200
days), which may affect the measurement of volatility spillovers.
Despite these limitations, the methodologies selected for this study were chosen
because they collectively offer a comprehensive framework for understanding the dynamic
interactions between strategic metals, precious metals, clean energy indices, EV
manufacturers, and traditional automakers. Each method addresses a specific aspect of these
complex relationships. DCC-GARCH (Engle, 2002b) captures time-varying correlations,
which is essential for examining how these relationships evolve during volatile periods like
46

the COVID-19 pandemic. VAR (Lütkepohl, 2005; C. A. Sims, 1980) models multi-directional
dependencies, essential for exploring how shocks propagate across asset classes, while Granger
Causality Tests (Granger, 1969) identify predictive relationships between variables. IRFs
visualize how shocks to one asset impact others over time, and the (Diebold & Yilmaz, 2012)
framework quantifies systemic risk and spillover effects. Therefore, these methods ensure a
detailed and dynamic analysis, aligning with the study’s goal of exploring interconnectedness,
risk transmission, and hedging potential in financial markets.
Chapter Four: EMPIRICAL RESULTS
This chapter presents the empirical results of this study, focusing on the dynamic
financial relationships between strategic metals, precious metals, and equity markets, including
clean energy indices, electric vehicle (EV) manufacturers, and traditional automakers. The
analysis begins with descriptive statistics to provide an initial understanding of the data
characteristics. An assessment of stationarity follows this to ensure that the time-series data is
suitable for advanced modeling.
The core analysis is conducted using several econometric techniques. First, the
Dynamic Conditional Correlation GARCH (DCC-GARCH) model is applied to examine timevarying correlations between metals and equity assets. Next, the Vector Autoregression (VAR)
model and Granger Causality tests explore the causal relationships between these variables,
identifying the directional influence of metals on stock returns. Impulse Response Functions
(IRF) are then used to illustrate how shocks in metal prices propagate through the financial
system.

47

Furthermore, the Total Connectedness Index (TCI), based on the Diebold-Yilmaz
(2012) methodology, quantifies volatility spillovers among the studied assets, highlighting
periods of increased market interconnectedness. The chapter concludes with an analysis of
hedge ratios and hedge effectiveness, comparing the performance of strategic and precious
metals as hedging instruments for clean energy and automotive stocks.
The findings of this chapter shed light on how strategic and precious metals behave in
relation to clean energy, electric vehicles, and traditional automotive markets. Understanding
these relationships can help investors and policymakers make informed decisions in
sustainability-focused financial markets.
4.1 Main results
Table 1 provides an overview of the descriptive statistics for strategic metals, precious
metals, EV manufacturers, clean energy indices, and traditional automakers. These statistics
offer insights into the characteristics of each asset, including their average performance,
volatility, and distribution.
The mean values indicate that Tesla has the highest average return (0.00115), reflecting
its rapid growth as a leading EV manufacturer. BYD also shows a positive average return
(0.00066), underscoring its strong position in the global EV market. These positive returns
suggest that the clean energy and EV sectors have experienced substantial growth during the
sample period, consistent with global sustainability trends (Cagli, 2023).
Among the strategic metals, all show positive average returns, including Copper
(0.00009), Aluminium (0.00013), Nickel (0.00005), and Zinc (0.00012). This performance
reflects their critical role in clean energy technologies and industrial applications (Alekseev et

48

al., 2024). This observation aligns with the increasing demand for critical minerals in the clean
energy transition (IEA, 2021).
The volatility (measured by standard deviation) is highest for Tesla (0.0351) and BYD
(0.0316), consistent with the inherent risk of high-growth technology stocks, which are
sensitive to market sentiment, technological advancements, and regulatory changes (Swarup
and Kushwaha, 2023). In contrast, Nickel (0.0223) and Palladium (0.0230) exhibit the highest
volatility among metals. These metals are known for their dual role as industrial inputs and
speculative financial assets, making them highly sensitive to market speculation and supply
chain disruptions (Swarup and Kushwaha, 2023).
The skewness and kurtosis values demonstrate that most return distributions are
negatively skewed and leptokurtic, indicating a higher likelihood of extreme negative returns.
This pattern is consistent with financial assets, which often experience extreme price
movements during periods of market stress (Cont, 2001). The kurtosis values exceed the
threshold of 3 for most variables, indicating leptokurtic distributions with heavy tails. Nickel
and Cobalt, in particular, show exceptionally high kurtosis, reflecting their sensitivity to supply
chain disruptions and market speculation. This observation is consistent with the findings of
(Swarup & Kushwaha, 2023) who highlighted that these metals are highly sensitive to market
shocks and price volatility due to their critical role in battery technologies and global supply
chains.
The results of the Jarque-Bera (JB) test for normality strongly reject the null hypothesis
(p < 0.01) for all variables, indicating that their return distributions deviate significantly from

49

normality. This is a common characteristic of financial time series, where returns often exhibit
non-normal behavior due to volatility clustering and extreme values (Brooks, 2019).
The Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests confirm the
stationarity of the return series, with all variables showing statistically significant results (p <
0.01). Stationarity is a crucial requirement for the application of advanced econometric models,
including DCC-GARCH, VAR, and Granger causality tests (Lütkepohl, 2005).
These descriptive statistics provide a foundation for the subsequent analysis, where
advanced econometric models will be used to explore the dynamic relationships between
strategic metals, precious metals, and various equity markets. This understanding is essential
for assessing the hedging potential of strategic metals and their role in clean energy and EV
sectors.
The correlation analysis presented in Fig. 1 reveals significant interrelationships among
strategic metals, electric vehicle (EV) manufacturers, clean energy indices, and traditional
automakers, offering insights into their interconnected dynamics.
Among strategic metals, copper exhibits a moderate positive correlation with BYD
(0.22) and a weaker positive correlation with Tesla (0.10). This relationship underscores
copper's critical role in EV manufacturing, particularly in battery systems and electrical
components (Nguyen et al., 2021). Nickel also shows a moderate positive correlation with
BYD (0.17), reflecting its importance in advanced battery technologies, especially for highperformance electric vehicles (Cagli, 2023). In contrast, cobalt displays weak correlations with
both Tesla (0.02) and BYD (0.05), suggesting that despite its significance in battery

50

technology, its price dynamics are less directly tied to EV manufacturers' performance,
possibly due to supply chain risks and market speculation (Bahini et al., 2024).
Table 1: Descriptive Statistics
Mean

Max

Min

SD

S

K

JB

PP

ADF

ECO

Variable

-0.00011

0.1339

-0.1625

0.0228

-0.2443

4.0772

1725.0069 ***

-48.8085 ***

-9.6776 ***

CELS

0.00028

0.1314

-0.1558

0.0209

-0.2861

4.7097

2303.041 ***

-50.0412 ***

-9.9168 ***

SPGTCLEN

0.00010

0.1103

-0.125

0.0151

-0.3554

8.1422

6837.6245 ***

-45.0941 ***

-15.0321 ***

Tesla

0.00115

0.2142

-0.2365

0.0351

-0.1001

4.8633

2424.0533 ***

-50.0007 ***

-11.4308 ***

BYD

0.00066

0.1857

-0.3402

0.0316

-0.1492

8.9806

8264.6705 ***

-49.558 ***

-22.2838 ***

Toyota

0.00039

0.1012

-0.0906

0.0153

0.116

3.3783

1172.7071 ***

-49.4071 ***

-15.6992 ***

Honda

0.00012

0.1009

-0.0868

0.0175

0.1917

3.2619

1103.195 ***

-50.4409 ***

-29.8668 ***

GM

0.00005

0.1818

-0.1902

0.0222

-0.0943

7.9836

6527.5353 ***

-49.2306 ***

-10.943 ***

F

-0.00007

0.2106

-0.1315

0.0221

0.1392

8.2954

7051.5127 ***

-48.7692 ***

-14.3147 ***

Copper

0.00009

0.0569

-0.0685

0.0127

-0.1425

2.0426

434.6127 ***

-52.0012 ***

-52.0584 ***

Aluminium

0.00013

0.0834

-0.1295

0.0137

-0.2884

6.0024

3721.0993 ***

-53.2528 ***

-16.3168 ***

Nickel

0.00005

0.3666

-0.3336

0.0223

0.4879

51.0247

266717.4174 ***

-54.3194 ***

-16.0465 ***

Zinc

0.00012

0.093

-0.0702

0.0161

0.106

1.551

250.2319 ***

-50.4946 ***

-50.4619 ***

Cobalt

-0.00003

0.1471

-0.291

0.0185

-2.8963

46.4634

224526.7257 ***

-55.3128 ***

-12.1721 ***

Gold

0.00021

0.0578

-0.0511

0.0094

0.0459

3.45

1218.1416 ***

-51.7727 ***

-19.9881 ***

Silver

0.00008

0.089

-0.1235

0.0179

-0.3908

5.5623

3228.6861 ***

-50.6884 ***

-20.0347 ***

Palladium

0.00019

0.226

-0.234

0.023

-0.2239

11.6642

13949.118 ***

-46.2638 ***

-20.4048 ***

Platinum

-0.00015

0.1118

-0.1232

0.0167

-0.1638

4.1678

1787.9952 ***

-49.1772 ***

-16.9039 ***

Note: Table 1 presents the descriptive statistics for the log returns of strategic metals, EV manufacturers, clean energy indices,
and market variables over the sample period. The Jarque-Bera (JB) test results indicate that most series deviate from normality
(p < 0.01). The Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests reject the presence of a unit root at the 1%
level, confirming stationarity for nearly all variables.
SD: Standard Deviation; S: Skewness; K: Kurtosis; JB: Jarque-Bera test; ADF: Augmented Dickey-Fuller test; PP: PhillipsPerron test; ADF and PP test for unit root (non-stationarity).
Significance: *** denotes 1%, ** denotes 5%, and ** denotes 10% significance.

The relationship between strategic metals and traditional automakers is notably weaker.
Copper's correlations with Toyota and Honda are both 0.13, while nickel's correlations with
these automakers are 0.08, indicating that traditional automakers, which rely more on internal
combustion engine (ICE) vehicles, are less affected by fluctuations in strategic metal prices.
This observation supports Hypothesis 2 (H2), suggesting that strategic metals have a greater
influence on the return volatility of EV stocks than on traditional automaker stocks.

51

Fig. 1 Correlation Heatmap of Strategic Metals, Precious Metals, EV Manufacturers, Traditional Automakers, and
Clean Energy Indices
Note: This figure visualizes the correlation structure among strategic metals, precious metals, EV manufacturers, traditional
automakers, and clean energy indices. Key positive and negative correlations highlight market interdependencies and potential
hedging opportunities, calculated using daily log returns over the sample period.

Strategic metals also show moderate positive correlations with clean energy indices,
including a correlation of 0.17 between copper and the Clean Energy Index (CELS) and 0.17
with ECO. This relationship highlights copper's role in renewable energy infrastructure, such
as solar panels, wind turbines, and energy storage systems (Fernanda Soares et al., 2025).
Nickel exhibits weaker correlations with these indices (0.08 with CELS and 0.09 with ECO).
In contrast, cobalt’s minimal correlations indicate that supply chain factors more influence its
price than clean energy demand (Bahini et al., 2024).
Precious metals such as gold and silver exhibit distinct behavior. Gold shows weak
correlations with EV stocks (0.05 with Tesla, 0.02 with BYD) and negative correlations with
52

traditional automakers (−0.07 with Toyota and Honda), underscoring its role as a safe-haven
asset (Coudert, V. and Raymond, H., 2011; S et al., 2022). Silver demonstrates a moderate
positive correlation with ECO (0.19), reflecting its dual role as both an industrial metal in solar
technology and a precious metal (Dutta, 2019).
These findings support Hypothesis 1 (H1), which suggests that significant time-varying
correlations exist between strategic metals, clean energy indices, EV stocks, and traditional
automaker stocks. They also provide preliminary evidence for Hypothesis 2 (H2), indicating
that strategic metals substantially influence EV stocks more than traditional automakers. In
subsequent sections, these initial insights will be further explored using advanced econometric
models.
Fig. 2 shows the dynamic relationships between strategic metals and four major sectors,
traditional automakers, electric vehicle (EV) manufacturers, clean energy indices, and precious
metals, revealing distinct interaction patterns. Among traditional automakers (Ford, GM,
Honda, and Toyota), the correlations with strategic metals, including copper, aluminum, and
nickel, are generally low but positive. These relationships are most pronounced for copper and
aluminum, reflecting their critical roles in vehicle manufacturing, electrical systems, and
lightweight components (Nguyen et al., 2021). However, cobalt consistently shows the
weakest correlation with traditional automakers, underscoring its primary relevance to battery
technologies rather than conventional internal combustion engine (ICE) vehicles (Bahini et al.,
2024). This low and unstable correlation of strategic metals aligns with Hypothesis 2 (H2),
suggesting that strategic metals have a greater influence on EV stocks than traditional
automaker stocks.

53

54

Fig. 2 Rolling Correlation between Strategic Metals and Equity Sectors (Automakers, EV Stocks, Clean Energy,
and Precious Metals)
Note: This figure illustrates the dynamic relationships between strategic metals (copper, aluminum, nickel, zinc, and cobalt)
and four major asset categories: (1) Traditional Automakers (Ford, GM, Honda, Toyota), (2) Electric Vehicle (EV) Stocks
(Tesla, BYD), (3) Clean Energy Indices (CELS, ECO, SPGTCLEN), and (4) Precious Metals (gold, silver, platinum,
palladium). The rolling correlation analysis (120-day window) highlights how these relationships evolve over time, with
notable shifts during global events such as the COVID-19 pandemic (2020) and the Russia-Ukraine conflict (2022). Positive
correlations indicate that metal prices and asset returns move together, while negative correlations suggest an inverse
relationship.

In contrast, the relationship between strategic metals and EV manufacturers (Tesla and
BYD) is significantly stronger, particularly for copper, aluminum, and nickel. These metals are
integral to battery technology, lightweight vehicle design, and electrical systems (Nguyen et
al., 2021). Tesla, which heavily relies on nickel-rich battery chemistries for its highperformance EVs, demonstrates a strong and consistent correlation with nickel (Cagli, 2023).
With its diversified battery strategy, BYD exhibits strong correlations with copper and
aluminum, consistent with its focus on lightweight, cost-effective EVs (Bahini et al., 2024).
During the COVID-19 pandemic and the Russia-Ukraine conflict, the correlations between EV
stocks and these strategic metals surged, highlighting their vulnerability to supply chain
disruptions and rising metal prices. These findings align with Hypothesis 1 (H1), which
suggests significant time-varying correlations between strategic metals and clean energy
equities, including EV stocks.
The clean energy sector, represented by indices such as CELS, ECO, and SPGTCLEN,
demonstrates strong and time-varying correlations with strategic metals, particularly copper
and aluminum. These metals are essential for renewable energy infrastructure, including solar
55

panels, wind turbines, and battery storage systems (Amândio et al., 2025). The rolling
correlation analysis reveals that these relationships intensified during global crises, such as the
COVID-19 pandemic and the Russia-Ukraine war, underscoring the sector's sensitivity to
commodity price fluctuations. In contrast, cobalt consistently shows weak and unstable
correlations with clean energy indices, reflecting its higher exposure to supply chain risks and
geopolitical factors rather than direct clean energy demand. This pattern supports Hypothesis
1 (H1) and Hypothesis 2 (H2), highlighting the critical role of strategic metals in clean energy
markets.
Analyzing precious metals (gold, silver, platinum, and palladium) against strategic
metals reveals distinct behaviors. Gold, known for its safe-haven properties, maintains low or
negative correlations with strategic metals, aligning with its role as a financial hedge during
market turmoil (Coudert, V. & Raymond, H., 2011). In contrast, silver exhibits stronger and
more volatile correlations with copper and nickel, reflecting its dual role as a precious and
industrial metal, particularly in solar technology (Dutta, 2019). Platinum and palladium
demonstrate moderate to high correlations with nickel and copper, particularly during periods
of market stress, such as the COVID-19 pandemic (Amândio et al., 2025). These findings align
with Hypothesis 5 (H5), which proposes that strategic metals offer greater hedging
effectiveness for clean energy and EV sectors than traditional precious metals.
Therefore, these findings provide a nuanced understanding of how strategic and
precious metals interact with different asset classes, underscoring their diverse roles as
industrial inputs, hedging tools, and safe-haven assets. The results reinforce the importance of
dynamic modeling approaches, such as rolling correlation analysis, to capture the evolving

56

relationships between metals and financial assets, especially in response to global economic
shocks and supply chain disruptions.
4.2 Volatility Connectedness Results (Diebold and Yilmaz, 2012)
The empirical findings of static volatility connectedness among strategic metals,
automakers (traditional and electric vehicles), clean energy indices, and precious metals are
presented in Fig. 3. This analysis employs the Diebold and Yilmaz (2012)time-domain
approach, providing a static view of how volatility is transmitted across these interconnected
markets. The Total Connectedness Index (TCI) is calculated at 49.4%, indicating a moderate
to high level of connectedness among markets throughout the entire sample period. This
suggests that volatility shocks in one market are likely to impact other markets within the
system.
The volatility spillover analysis reveals differences between Tesla and BYD regarding
their connections with strategic metals. Tesla exhibits a relatively high spillover from cobalt
(2.27%) and nickel (2.30%), highlighting the critical importance of these metals in its battery
technologies. These findings align with Tesla’s focus on high-performance batteries, where
nickel and cobalt are essential for enhancing energy density and battery life (Graedel et al.,
2015). In contrast, BYD shows a broader sensitivity to strategic metals, including Nickel
(2.63%) and aluminum (2.65%), which reflects its diversified manufacturing strategy,
encompassing EVs, batteries, and renewable energy technologies (IEA, 2023). This higher
spillover from copper emphasizes BYD’s extensive use of this metal in electrical components
and lightweight vehicle structures, which is critical for its cost-effective EV models (Nguyen
et al., 2021). These results support Hypothesis 1 (H1), indicating that strategic metals have a
more significant influence on EV stocks than on traditional automakers.
57

Tesla and BYD exhibit strong volatility connections with clean energy indices,
particularly CELS (9.42% for Tesla and 2.72% for BYD). This robust connection suggests that

Fig. 3 Static Volatility Spillovers Among Strategic Metals, Automakers, Clean Energy, and Precious Metals
(2014–2024)
Note: This figure displays the static volatility spillover matrix among strategic metals (Copper, Nickel, Cobalt), automakers
(EVs: Tesla, BYD; Traditional: Toyota, GM), clean energy indices (CELS, ECO, SPGTCLEN), and precious metals (Gold,
Silver, Platinum, Palladium) for the period 2014–2024. The Total Connectedness Index (TCI) of 49.4% indicates moderateto-high systemic interdependence, with clean energy indices (CELS, ECO) acting as major volatility receivers and strategic
metals (Copper, Nickel) as dominant transmitters. Net spillovers highlight Tesla/BYD’s sensitivity to battery metals (Cobalt,
Nickel) and GM/Ford’s dual role as transmitters to clean energy indices.

the performance of these EV manufacturers is closely tied to the broader clean energy sector,
where policy changes, technological advancements, and global sustainability trends are key
58

drivers (Bastianin et al., 2025). Tesla’s higher spillover from CELS indicates that clean energy
trends and investor sentiment more directly influence its stock price. In contrast, BYD’s more
balanced exposure to clean energy indices (including SPGTCLEN and ECO) reflects its
diversified approach, which includes EV production, battery manufacturing, and renewable
energy solutions. The relatively lower spillover from SPGTCLEN (1.64% for Tesla and 2.22%
for BYD) suggests that this index, with its broader range of clean energy companies, is less
directly connected to the specific dynamics of EV manufacturing. These findings support
Hypothesis 1 (H1), which proposes significant time-varying correlations between strategic
metals and clean energy indices.
Tesla and BYD demonstrate weaker volatility spillovers from precious metals,
reflecting the limited role of these metals in their core operations. For Tesla, gold (1.67%) and
silver (1.53%) exhibit the lowest spillovers, consistent with their status as safe-haven assets
rather than industrial inputs (Baur & Lucey, 2010b). In contrast, platinum and palladium show
slightly higher spillovers (2.23% and 2.33% for Tesla), which can be attributed to their
industrial applications, particularly in catalytic converters. BYD’s slightly higher sensitivity to
silver (2.16%) and platinum (1.88%) may be linked to its diversified production, which
includes solar energy solutions where silver is a critical component (Dutta, 2019). These results
support Hypothesis 5 (H5), indicating that precious metals, especially gold and silver, provide
greater hedging effectiveness for automotive stocks compared to non-precious metals.
Traditional automakers (Toyota, Honda, GM, and Ford) show stronger volatility
spillovers from copper (3.58% for GM, 2.68% for Toyota) and aluminum (2.23% for GM,
2.22% for Toyota), highlighting their dependence on these metals for internal combustion
engine (ICE) vehicles. These metals are critical for producing electrical systems, engine
59

components, and lightweight body structures (Graedel et al., 2015). However, cobalt shows
minimal spillover to traditional automakers (1.80% for Toyota), emphasizing its primary
association with battery production rather than conventional vehicle manufacturing. These
findings align with Hypothesis 2 (H2), suggesting that strategic metals, particularly those
associated with EV battery technology, have a more significant impact on EV manufacturers
than on traditional automakers.
Traditional automakers exhibit moderate spillovers from clean energy indices, with
GM (6.61% from CELS) and Ford (6.35% from CELS) receiving the highest spillovers. This
finding suggests that while these companies are transitioning toward sustainable technologies,
their performance is still primarily driven by traditional automotive dynamics (Graedel et al.,
2015). In contrast, Toyota (3.08% from CELS) and Honda (2.50% from CELS) demonstrate
lower sensitivity to clean energy indices, reflecting their focus on hybrid vehicles and internal
combustion engine (ICE) models. These results support Hypothesis 1 (H1), which proposes
that significant time-varying correlations exist between strategic metals, clean energy indices,
and automotive stocks.
Traditional automakers show moderate spillovers from precious metals, particularly
gold (2.44% for GM, 1.46% for Toyota) and silver (2.07% for GM, 1.45% for Toyota). These
metals serve as safe-haven assets, providing stability during market volatility but are not
directly connected to automotive manufacturing (Baur & McDermott, 2010). Platinum and
palladium exhibit slightly higher spillovers (2.60% for Toyota, 3.13% for GM), consistent with
their industrial use in catalytic converters for ICE vehicles (Dutta, 2019). These results support
Hypothesis 5 (H5), suggesting that while precious metals can offer hedging benefits, they are
less directly tied to traditional automakers’ core operations.
60

The net spillover (To others − From others) analysis provides further insights into the
dynamic interactions among these assets. Clean energy indices (CELS: 19.2%, ECO: 14%) are
the largest net transmitter of volatility, reflecting their high sensitivity to market-wide shocks,
while SPGTCLEN (-5.3%) shows lower volatility absorption due to its diversified portfolio of
clean energy companies. In contrast, GM (10.5%) and Ford (10.8%) are the largest net
transmitters of volatility among automakers, highlighting their significant influence on the
broader market, possibly due to their scale and global operations. Tesla (-2.52%) and BYD (5.92%) are net receivers of volatility, underscoring their sensitivity to external shocks,
particularly those related to strategic metals used in battery production (Lander et al., 2021).
Precious metals show distinct net spillover patterns, with gold (1.25%) acting as a mild
transmitter of volatility, while silver (-0.71%), platinum (-0.86%), and palladium (-5.33%) are
net receivers. Palladium, in particular, is the most significant net receiver, consistent with its
reliance on industrial demand in the automotive sector (Erb and Harvey, 2006). These results
provide partial support for Hypothesis 5 (H5), suggesting that while strategic metals like nickel
and cobalt play a key role in volatility transmission to EV and clean energy stocks, precious
metals, particularly gold, can also offer comparable or even stronger hedging potential in
certain contexts.
Fig. 4 illustrates the Total Connectedness Index (TCI) from January 30, 2014, to
January 30, 2024, reflecting the dynamic volatility spillovers among strategic metals, electric
vehicle (EV) manufacturers, clean energy indices, and precious metals. Prior to 2020, TCI
values fluctuated within a relatively stable range of 40.43% to 53.83%, indicating moderate
interconnectedness and limited systemic risk under normal market conditions.

61

Fig. 4 Total Connectedness Index (TCI) from Jan 30, 2014, to Jan 30, 2024
Note: This figure shows the time-varying volatility spillovers among strategic metals, EV manufacturers, clean energy indices,
and precious metals. Sharp spikes during March 2020 and March 2022 reflect intensified systemic risk following the COVID19 outbreak and the Russia–Ukraine conflict, respectively.

A sharp surge in TCI is observed in March 2020, coinciding with the World Health
Organization's declaration of COVID-19 as a global pandemic. (World Health Organization,
2020). Specifically, on March 12, 2020, the TCI reached 90.67%, and it peaked at 94.44% on
March 16 and 18, 2020. This surge reflects heightened market interconnectedness driven by
global uncertainty during the pandemic's onset. These findings align with the study by Tan et
al. (2022) , which reported that the COVID-19 outbreak significantly increased the total
volatility spillover level among energy and precious metals markets, enhancing risk
connectivity.
Post-2020, TCI experiences a significant decline, followed by a stabilization phase.
However, another notable increase occurred in early March 2022, aligning with the onset of
the Russia–Ukraine conflict. On March 7 and 8, 2022, the TCI again reached 94.44%,
indicating a resurgence in market interconnectedness due to geopolitical tensions. This
observation is supported by the research of Xing et al. (2023), who found that the Russia–
62

Ukraine conflict significantly impacted the risk transmission among energy subsector stocks
in China, altering the volatility correlations during the conflict period. These observations
emphasize the strong interdependence among sectors, particularly during periods of global
crises.
4.3 Dynamic Conditional Correlation GARCH (DCC-GARCH) Results
The empirical results of the Dynamic Conditional Correlation GARCH (DCCGARCH) model, as presented in Table 2, provide a comprehensive view of how the dynamic
correlations between strategic metals, precious metals, and various equity markets, such as
traditional automakers, electric vehicle (EV) manufacturers, and clean energy indices, evolve
over time. This model is particularly useful for capturing the time-varying nature of
correlations, making it suitable for analyzing how these relationships fluctuate in response to
market conditions, economic events, and global crises.
The DCC-GARCH model is characterized by two key parameters: the α (alpha) and β
(beta) coefficients. The alpha coefficient measures the short-term impact of shocks on the
dynamic correlations between the asset pairs. A significant and positive alpha indicates that
recent market shocks have a strong influence on the correlation between the two assets. In
contrast, the beta coefficient reflects the persistence of these correlations over time, capturing
their long-term stability. A high sum of α and β (close to 1) indicates strong persistence in the
dynamic correlations, suggesting that once volatility emerges, it tends to remain elevated, a
phenomenon known as volatility clustering (Engle, 2002b).
To ensure the robustness of the DCC-GARCH model, the optimal values of p (AR
order) and q (MA order) for each pair of assets (metals and equity index) were determined
using the maximum values of p and q derived from the ARMA models of each series. This
63

approach guarantees that the model captures the most complex time-series patterns of both
series, avoiding underfitting. By selecting the highest p and q values among the two series
being analyzed, we ensured that the model could account for both short-term fluctuations and
long-term dependencies in each pair. This methodological decision enhances the model's
ability to fully capture the autocorrelation and moving average patterns in both series.
The DCC-GARCH analysis provides a comprehensive understanding of the dynamic
relationships between metals (strategic (non-precious) and precious metals) and various equity
markets, including traditional automakers, electric vehicle (EV) manufacturers, and clean
energy indices. The results reveal distinct patterns of correlation persistence (β values) and
short-term shock sensitivity (α values) across different asset groups, aligning with the study's
hypotheses.
The results for traditional automakers (Toyota, Honda, GM, and Ford) paired with nonprecious metals (copper, aluminum, nickel, zinc, and cobalt) indicate that most metalautomaker pairs exhibit high persistence in their correlations, with β values approaching 1
(e.g., Toyota-Copper: β = 0.9749, Honda-Copper: β = 0.9926). This strong persistence is
consistent with the structural importance of copper and aluminum in automotive
manufacturing, including electrical systems and lightweight materials (Graedel et al., 2015).
However, cobalt shows significantly lower persistence (GM-Cobalt: β = 0.574), reflecting its
primary association with battery technologies rather than traditional automotive components
(van den Brink et al., 2020). These findings support Hypothesis 2 (H2), which suggests that
strategic metals have a greater influence on EV stocks than on traditional automaker stocks.
The DCC-GARCH analysis for automakers (Toyota, Honda, GM, and Ford) paired
with precious metals (gold, silver, platinum, and palladium) reveals diverse patterns in their
64

dynamic correlations. Most of the β coefficients are statistically significant and high (e.g., β =
0.9552 for GM-Gold, β = 0.9726 for F-Gold), indicating strong long-term persistence in their
correlations. This stability aligns with the concept of volatility clustering1, where past volatility
influences future volatility (Engle, 2002b). Precious metals, traditionally viewed as safe-haven
assets (Baur and Lucey, 2010b; Coudert, V. and Raymond, H., 2011), demonstrate relatively
high persistence with automakers, particularly for gold and silver. However, the α coefficients,
representing short-term shock sensitivity, are generally low, suggesting that while these
correlations are stable, they are not highly reactive to immediate market events. This
characteristic is consistent with the safe-haven behavior of precious metals, where their value
tends to maintain stability during market turbulence (Baur and McDermott, 2010; Caporale
and Gil-Alana, 2023). The weaker α values for gold and silver (e.g., α = 0.0148 for ToyotaGold, α = 0.0075 for Honda-Silver) reinforce this observation, indicating that short-term
shocks have minimal impact on their relationship with automakers. These findings support
Hypothesis 5 (H5) of this study, which proposes that precious metals, especially gold and
silver, offer greater hedging effectiveness for automotive stocks compared to non-precious
metals.
Table 2: Dynamic Conditional Correlation (DCC-GARCH) Results for Automakers, EV Manufacturers, and
Clean Energy Indices Paired with Metals
Pair of assets

ARMA (p, q)

DCC(α)

DCC(β)

Toyota and Copper

ARMA(3,2)

0.004

0.9749 ***

Toyota and Aluminium

ARMA(3,2)

0.0035

0.9794 ***

Toyota and Nickel

ARMA(3,2)

0.0121 *

0.9447 ***

Toyota and Zinc

ARMA(3,2)

1e-04

0.7653

Toyota and Cobalt

ARMA(3,2)

0.0426 ***

0.8495 ***

1

Volatility clustering is a phenomenon in financial time series where periods of high volatility tend to be followed
by high volatility, and periods of low volatility are followed by low volatility. This characteristic is commonly
observed in asset returns and is a key concept in financial econometrics, particularly in GARCH-type models
(Engle, 2002b).

65

Toyota and Gold

ARMA(3,3)

0.0148

0.8887 ***

Toyota and Silver

ARMA(3,2)

0.0025

0.9016 ***

Toyota and Platinum

ARMA(3,2)

0.0142

0.821 ***

Toyota and Palladium

ARMA(3,2)

0.0102

0.9181 ***

Honda and Copper

ARMA(3,1)

0.0031

0.9926 ***

Honda and Aluminium

ARMA(3,2)

0

0.9256

Honda and Nickel

ARMA(3,1)

0.0066

0.9576 ***

Honda and Zinc

ARMA(3,0)

0

0.894 ***

Honda and Cobalt

ARMA(3,2)

0.0038

0.9128 ***

Honda and Gold

ARMA(3,3)

0.0105

0.9288 ***

Honda and Silver

ARMA(3,0)

0.0075 *

0.9486 ***

Honda and Platinum

ARMA(3,0)

0.0217 *

0.7825 ***

Honda and Palladium

ARMA(3,1)

0.0138

0.9009 ***

GM and Copper

ARMA(0,1)

0.0127 **

0.9654 ***

GM and Aluminium

ARMA(2,2)

0.0068

0.9549 ***

GM and Nickel

ARMA(0,1)

0.0119 **

0.9716 ***

GM and Zinc

ARMA(0,0)

0.0095 *

0.9711 ***

GM and Cobalt

ARMA(2,2)

0.0344

0.574 ***

GM and Gold

ARMA(1,3)

0.0206 ***

0.9552 ***

GM and Silver

ARMA(0,0)

0.0151 **

0.9545 ***

GM and Platinum

ARMA(0,0)

0.0166 **

0.9704 ***

GM and Palladium

ARMA(0,1)

0

0.9584

F and Copper

ARMA(0,1)

0.0246 **

0.8235 ***

F and Aluminium

ARMA(2,2)

0.0038

0.8474 ***

F and Nickel

ARMA(0,1)

0.0173 *

0.883 ***

F and Zinc

ARMA(0,0)

0.0133 *

0.9157 ***

F and Cobalt

ARMA(2,2)

0.0197

0.674 ***

F and Gold

ARMA(1,3)

0.0133 **

0.9726 ***

F and Silver

ARMA(0,0)

0.0143 **

0.958 ***

F and Platinum

ARMA(0,0)

0.0146 *

0.9675 ***

F and Palladium

ARMA(0,1)

0.0071

0.9699 ***

Tesla and Copper

ARMA(0,1)

0.0088

0.761 ***

Tesla and Aluminium

ARMA(2,2)

0.018

0.7661 ***

Tesla and Nickel

ARMA(0,1)

0.0098

0.8833 ***

Tesla and Zinc

ARMA(0,0)

0.0249

0.5654 *

Tesla and Cobalt

ARMA(2,2)

0.0057 *

0.9676 ***

Tesla and Gold

ARMA(1,3)

0.0122

0.8801 ***

Tesla and Silver

ARMA(0,0)

0.0096

0.8592 ***

Tesla and Platinum

ARMA(0,0)

0.011

0.8851 ***

Tesla and Palladium

ARMA(0,1)

0

0.9355 *

BYD and Copper

ARMA(0,1)

0.0124 **

0.9762 ***

BYD and Aluminium

ARMA(2,2)

8e-04

0.868 ***

BYD and Nickel

ARMA(0,1)

0.0148 *

0.9113 ***

BYD and Zinc

ARMA(0,0)

0.002

0.9866 ***

66

BYD and Cobalt

ARMA(2,2)

0.1707 ***

0.4763 ***

BYD and Gold

ARMA(1,3)

0.0063

0.9292 ***

BYD and Silver

ARMA(0,0)

0.0148

0.8825 ***

BYD and Platinum

ARMA(0,0)

0.015

0.8875 ***

BYD and Palladium

ARMA(0,1)

0.007

0.9303 ***

CELS and Copper

ARMA(4,4)

0.0103 **

0.9749 ***

CELS and Aluminium

ARMA(4,4)

0.0084

0.9299 ***

CELS and Nickel

ARMA(4,4)

0.0463 ***

0.6753 ***

CELS and Zinc

ARMA(4,4)

0.0073

0.8864 ***

CELS and Cobalt

ARMA(4,4)

0.0036

0.9817 ***

CELS and Gold

ARMA(4,4)

0.0321 ***

0.9468 ***

CELS and Silver

ARMA(4,4)

0.0218 **

0.949 ***

CELS and Platinum

ARMA(4,4)

0.0199 ***

0.967 ***

CELS and Palladium

ARMA(4,4)

2e-04

0.9776 ***

SPGTCLEN and Copper

ARMA(0,2)

0.0179

0.1193

SPGTCLEN and Aluminium

ARMA(2,2)

0

0.9311 *

SPGTCLEN and Nickel

ARMA(0,2)

0.0015

0.992 ***

SPGTCLEN and Zinc

ARMA(0,2)

0.0015

0.9894 ***

SPGTCLEN and Cobalt

ARMA(2,2)

0.0126 *

0.9374 ***

SPGTCLEN and Gold

ARMA(1,3)

0.0394 *

0.4801 ***

SPGTCLEN and Silver

ARMA(0,2)

0.0692 **

0.236

SPGTCLEN and Platinum

ARMA(0,2)

0.0379 **

0.5201 ***

SPGTCLEN and Palladium

ARMA(0,2)

0

0.9224 ***

ECO and Copper

ARMA(1,2)

0.014 ***

0.9679 ***

ECO and Aluminium

ARMA(2,2)

0.0101 *

0.9587 ***

ECO and Nickel

ARMA(1,2)

0.0349 ***

0.7815 ***

ECO and Zinc

ARMA(1,2)

0.0083

0.8931 ***

ECO and Cobalt

ARMA(2,2)

0.0024

0.9812 ***

ECO and Gold

ARMA(1,3)

0.028 ***

0.9497 ***

ECO and Silver

ARMA(1,2)

0.0136 *

0.9654 ***

ECO and Platinum

ARMA(1,2)

0.0165 ***

0.9712 ***

ECO and Palladium

ARMA(1,2)

0

0.9732 ***

Note: This table presents the results of the DCC-GARCH model, showing the α (short-term shock impact) and β (long-term
persistence) coefficients for the dynamic correlations between automakers, EV manufacturers, clean energy indices, and
both non-precious (copper, aluminum, nickel, zinc, cobalt) and precious metals (gold, silver, platinum, palladium).
Significant coefficients are marked with ***, **, and * for 1%, 5%, and 10% significance levels. A high sum of α + β (close
to 1) indicates strong persistence and volatility clustering.

The DCC-GARCH analysis for Tesla and BYD paired with both non-precious (copper,
aluminum, nickel, zinc, and cobalt) and precious metals (gold, silver, platinum, and palladium)
reveals distinct and meaningful patterns. Among the non-precious metals, Tesla's correlations
show strong persistence (high β values), particularly with cobalt (β = 0.968) and nickel (β =

67

0.883), which are critical for battery technology in EV production (Lander et al., 2021).
However, the α coefficients, which capture short-term sensitivity to market shocks, are
generally low, with the highest being for zinc (α = 0.025) and the lowest for cobalt (α = 0.006).
This indicates that while Tesla’s relationships with these metals are stable over time, they are
less reactive to immediate market changes.
In contrast, BYD's correlations with these metals exhibit a more varied pattern. Cobalt
demonstrates a high α value (α = 0.171) but a relatively low β (β = 0.476), suggesting a strong
sensitivity to short-term shocks but limited long-term stability. This aligns with the volatile
nature of the cobalt market, where supply chain risks and geopolitical factors heavily influence
prices (van den Brink et al., 2020). On the other hand, copper and zinc show high persistence
(β = 0.976 and 0.987, respectively), reflecting their fundamental role in EV manufacturing.
These results support Hypothesis 1 (H1), which suggests significant time-varying correlations
between strategic metals and EV stocks.
The results are more stable when examining the relationship between EV
manufacturers and precious metals. Tesla's correlations with gold, silver, platinum, and
palladium are characterized by high β values, indicating long-term persistence, but low α
values, suggesting minimal sensitivity to short-term shocks. For instance, Tesla’s correlation
with gold shows a high β of 0.880 but a low α of 0.012, consistent with the traditional view of
gold as a safe-haven asset (Baur and Lucey, 2010a). Similarly, BYD demonstrates low α values
for all precious metals, with the lowest for gold (α = 0.006) and the highest for silver and
platinum (α = 0.015). These findings reinforce the perception of precious metals as stable, lowvolatility assets that provide hedging benefits for EV stocks, aligning with Hypothesis 5 (H5).

68

The DCC-GARCH analysis for clean energy indices (CELS, SPGTCLEN, and ECO)
paired with various non-precious metals (copper, aluminum, nickel, zinc, and cobalt) and
precious metals (gold, silver, platinum, and palladium) reveal significant differences in their
dynamic correlations. The results demonstrate that non-precious metals, particularly copper
and cobalt, exhibit high long-term persistence (β values) in their correlations with clean energy
indices. For example, CELS and ECO maintain strong long-term relationships with copper (β
= 0.9749 for CELS, 0.9679 for ECO) and cobalt (β = 0.9817 for CELS, 0.9812 for ECO),
reflecting their critical roles in renewable energy technologies, including solar panels, wind
turbines, and battery systems (Nguyen et al., 2021). This high persistence aligns with
Hypothesis 1 (H1), which suggests that strategic metals, particularly those linked to clean
energy production, demonstrate strong time-varying correlations with clean energy indices
(Amândio et al., 2025). In contrast, nickel, while critical for battery technologies, demonstrates
moderate short-term sensitivity (α = 0.046 for CELS, α = 0.035 for ECO), reflecting its
exposure to supply chain disruptions and market speculation (Erb and Harvey, 2006). This
short-term volatility is consistent with the risks associated with battery metals in global
markets.
The DCC-GARCH analysis of clean energy indices (CELS, SPGTCLEN, and ECO)
paired with strategic metals (copper, aluminum, nickel, zinc, and cobalt) and precious metals
(gold, silver, platinum, and palladium) reveal significant insights into their dynamic
relationships. The results demonstrate that strategic metals, particularly copper and cobalt,
exhibit strong and persistent correlations (high β values) with CELS and ECO. For instance,
the correlation between CELS and copper (β = 0.9749) and cobalt (β = 0.9817) highlights the
critical role of these metals in renewable energy technologies, such as solar panels, wind
69

turbines, and energy storage systems (Fernanda Soares et al., 2025). These high persistence
values suggest that clean energy indices are directly influenced by the supply and demand
dynamics of strategic metals, making them essential components for clean technology
development. This finding supports Hypothesis 1 (H1) of the study, which proposes that
significant time-varying correlations exist between strategic metals and clean energy indices.
Conversely, the analysis of precious metals, particularly gold and silver, reveals weaker
and less stable correlations with clean energy indices, especially SPGTCLEN. For
SPGTCLEN, gold (β = 0.480) and silver (β = 0.236) exhibit the lowest persistence among the
indices, reflecting their limited direct application in clean energy technologies (Baur & Lucey,
2010b). These results suggest that, while precious metals like gold are traditionally considered
safe-haven assets (Baur & McDermott, 2010), they do not provide effective hedging benefits
for clean energy investments. Silver, which has dual properties as a precious and industrial
metal, shows moderate persistence with CELS (β = 0.949) and ECO (β = 0.9654), reflecting
its role in solar technology (Dutta, 2019). These findings align with Hypothesis 5 (H5), which
proposes that strategic metals offer superior hedging potential for clean energy stocks
compared to precious metals.
To visualize the time-varying relationships captured by the DCC-GARCH model,
correlation and covariance plots were generated for all pairs of strategic and precious metals
with equity sectors, including electric vehicle (EV) manufacturers, traditional automakers, and
clean energy indices. These multi-panel plots (Fig. 5) incorporate key global events, such as
the COVID-19 outbreak (March 2020) and the Russia–Ukraine conflict (March 2022), to
facilitate crisis-based interpretation. This visualization offers a comprehensive overview of

70

dynamic co-movement patterns, reinforcing the temporal insights derived from the model
estimates.
The dynamic covariance analysis reveals notable differences in how Tesla and BYD interact
with both Strategic and Precious Metals across major crisis periods. In 2020, Tesla exhibited
strong upward covariance jumps with most metals, indicating elevated co-movement during
the COVID-19 shock. In contrast, the 2022 response was more muted and stable, with the
Tesla–Nickel pair briefly dropping into negative territory before reverting to near-zero levels.
For BYD, Aluminum and Nickel showed limited reaction in 2020, while other Strategic Metals
such as Copper, Cobalt, and Zinc displayed sharp positive spikes. All Precious Metals,
including Gold, Silver, Palladium, and Platinum, experienced positive covariance jumps with
BYD during this period. However, in 2022, BYD’s responses were more heterogeneous:
Aluminum, Zinc, Palladium, and Platinum showed positive reactions; Copper rose modestly;
Nickel dropped negatively; and Gold and Silver experienced slight negative shifts. Overall,
since most covariances were near or slightly above zero and exhibited temporary jumps during
global stress periods, these metals lack consistent hedging capacity under normal conditions.
This finding aligns with earlier studies suggesting that asset co-movement tends to rise during
crises but does not persist structurally over time (Baur & Lucey, 2010b; Dutta et al., 2020;
Lahiani et al., 2021).

71

72

73

74

75

Fig. 5 Comparative Time Series of Rolling Conditional Correlations and Covariances between Equity Sectors
and Metal Markets.
Note: This figure presents the rolling conditional correlation and covariance dynamics between traditional automakers, EV
manufacturers, and clean energy indices in relation to both Precious and Strategic (non-precious) Metals. The analysis
highlights how these relationships evolve over time, particularly around major global events such as the COVID-19 pandemic
and the Russia–Ukraine conflict.

In contrast, traditional automakers, including Toyota, Honda, GM, and Ford, exhibit
more stable but generally weaker correlations with non-precious metals such as Aluminium
and Copper. These metals are widely used in structural components, wiring systems, and
lightweight vehicle frames, supporting their continued functional relevance across both
internal combustion engine (ICE) and hybrid platform(IEA, 2021b). Copper displays stronger
co-movement during systemic disruptions, such as the COVID-19 pandemic and the Russia–
Ukraine conflict, likely due to its centrality in global supply chains. However, correlations with
battery-specific metals like Cobalt and Nickel remain low and more volatile, suggesting limited

76

dependence on energy storage technologies compared to EV manufacturers (Gustafsson et al.,
2022). Therefore, these findings support Hypothesis 2, showing that strategic metals influence
traditional automakers less than EV manufacturers.
Clean energy indices—ECO, CELS, and SPGTCLEN—exhibit moderate to strong,
time-varying correlations with Strategic Metals, particularly Copper and Aluminium. These
correlations intensify during periods of global stress, underscoring the critical dependence of
renewable technologies on these metals for infrastructure such as wind turbines, solar panels,
and energy storage systems (Sovacool et al., 2020; IEA, 2021b). Among the indices, ECO
shows the most consistent and persistent positive correlation with Copper, while CELS and
SPGTCLEN display more volatility. In contrast, correlations with Cobalt and Nickel are
generally weaker and less stable, reflecting the complexity of their supply chains and exposure
to geopolitical risk (Erdoğan et al., 2022). These patterns support the idea that strategic metals
play a key role in clean energy market behavior, especially during global disruptions.
Therefore, these findings support Hypothesis 2 by confirming that Strategic Metals
have a more limited and less persistent financial influence on traditional automakers compared
to electric vehicle manufacturers.
Table 3 presents summary statistics of the dynamic conditional correlations between
equity sectors, including EV manufacturers, traditional automakers, and clean energy indices,
and both Strategic and Precious Metals. The results reveal clear sectoral differences in the
magnitude and stability of correlations. EV manufacturers such as BYD and Tesla exhibit
generally higher and more volatile average correlations with Strategic Metals, particularly
Copper, Nickel, and Zinc, reflecting their stronger integration with mineral supply chains
essential to battery and vehicle production (IEA, 2021). BYD shows higher mean correlations
77

with Copper (0.20) and Zinc (0.18) than Tesla, which maintains more moderate values. In
contrast, traditional automakers like Toyota, Honda, GM, and Ford demonstrate weaker and
more stable correlations, especially with battery metals such as Cobalt and Nickel,
underscoring their limited reliance on energy storage technologies (Gustafsson et al., 2022).
Clean energy indices show the strongest alignment with Strategic Metals among all groups.
ECO and CELS record consistently high mean correlations with Copper and Platinum, while
SPGTCLEN exhibits weak or even negative correlations, likely due to its broader and more
diversified composition (Sovacool et al., 2020). Across all sectors, Precious Metals such as
Gold and Silver show weak and inconsistent correlations, further supporting their classification
as safe-haven assets rather than sector-linked instruments (Baur & Lucey, 2010b). Overall,
these findings support Hypotheses 1 and 2 by confirming that Strategic Metals are more
connected to EV and clean energy sectors than to traditional automakers, while Precious Metals
behave more as safe-haven assets with limited hedging capacity.
The diagnostic tests in Table 4 exhibit a comprehensive evaluation of the ARCH effects
and the Ljung–Box Q-statistic for the series examined. Consistently across the series, the
ARCH LM test at both 5 and 10 lags reveals highly significant values (p-value < 0.01),
suggesting a robust presence of conditional heteroscedasticity and validating the application
of GARCH-type models to capture the observed volatility patterns.
The Ljung–Box Q-statistic is used to detect autocorrelation in the residuals. While
several series such as CELS, SPGTCLEN, and ECO exhibit statistically significant
autocorrelation, other series, such as Tesla, Honda, Zinc, and Silver, do not indicate that their
model residuals resemble white noise. This suggests that the ARMA components adequately
captured their short-term dynamics.
78

Table 3 Statistical Summary of Dynamic Conditional Correlations Between equities and metals
Pair

Mean

Median

SD

Min

Max

Toyota - Copper

0.1091

0.1106

0.0223

-0.0075

0.1658

Toyota - Aluminium

0.0815

0.0815

0.0197

0.0305

0.1416

Toyota - Nickel

0.0853

0.0868

0.0461

-0.2538

0.2673

Toyota - Zinc

0.1032

0.1032

0.0002

0.1016

0.1039

Toyota - Cobalt

0.0405

0.0383

0.0371

-0.1218

0.2958

Toyota - Gold

-0.0829

-0.0815

0.0368

-0.3644

0.1069

Toyota - Silver

-0.0175

-0.0178

0.0062

-0.0402

0.0262

Toyota - Platinum

0.0136

0.0132

0.0267

-0.0923

0.1434

Toyota - Palladium

0.0576

0.0595

0.0299

-0.0443

0.2159

Honda - Copper

0.1178

0.1149

0.0392

0.0292

0.2059

Honda - Aluminium

0.0797

0.0797

0

0.0797

0.0797

Honda - Nickel

0.0949

0.0967

0.0267

-0.089

0.1756

Honda - Zinc

0.1022

0.1022

0

0.1022

0.1022

Honda - Cobalt

0.0108

0.0105

0.0094

-0.0754

0.0936

Honda - Gold

-0.0795

-0.0791

0.0351

-0.2724

0.0944

Honda - Silver

-0.0123

-0.014

0.0283

-0.099

0.1788

Honda - Platinum

0.0229

0.0219

0.0382

-0.1542

0.2079

Honda - Palladium

0.0639

0.0654

0.0363

-0.1171

0.2901

GM - Copper

0.1398

0.1348

0.0618

-0.0763

0.496

GM - Aluminium

0.0601

0.0582

0.0263

-0.0313

0.2247

GM - Nickel

0.0759

0.0763

0.0662

-0.1797

0.3697

GM - Zinc

0.079

0.0778

0.0483

-0.0647

0.3222

GM - Cobalt

-0.0031

-0.0031

0.0246

-0.33

0.3728

GM - Gold

-0.0595

-0.0592

0.091

-0.3614

0.2286

GM - Silver

0.0591

0.0579

0.0608

-0.153

0.2609

GM - Platinum

0.1428

0.1524

0.1034

-0.1724

0.4249

GM - Palladium

0.1443

0.1443

0

0.1443

0.1443

F - Copper

0.1184

0.1183

0.0447

-0.0743

0.4615

F - Aluminium

0.0686

0.0683

0.0072

0.0274

0.1272

F - Nickel

0.0745

0.0744

0.0403

-0.1415

0.3492

F - Zinc

0.0858

0.0861

0.0358

-0.0513

0.2938

F - Cobalt

0.0022

0.0021

0.0174

-0.1829

0.3155

F - Gold

-0.0548

-0.0534

0.0819

-0.3089

0.1858

F - Silver

0.0542

0.0514

0.0577

-0.1171

0.2601

F - Platinum

0.1295

0.1311

0.0806

-0.1411

0.3577

F - Palladium

0.1286

0.1294

0.0352

0.0305

0.2383

Tesla - Copper

0.0766

0.0771

0.0147

-0.0017

0.2084

Tesla - Aluminium

0.0337

0.0333

0.0285

-0.1595

0.1863

Tesla - Nickel

0.069

0.0683

0.0221

-0.0593

0.2042

Tesla - Zinc

0.0751

0.0755

0.0279

-0.1121

0.2381

Tesla - Cobalt

0.0297

0.0283

0.019

-0.0344

0.131

79

Tesla - Gold

0.0159

0.0145

0.0266

-0.1091

0.1458

Tesla - Silver

0.0721

0.072

0.0202

-0.0456

0.2327

Tesla - Platinum

0.1201

0.1202

0.0246

0.0371

0.314

Tesla - Palladium

0.1224

0.1224

0

0.1224

0.1224

BYD - Copper

0.2004

0.2035

0.0805

-0.1134

0.3991

BYD - Aluminium

0.1401

0.1401

0.0018

0.1289

0.1477

BYD - Nickel

0.1746

0.1733

0.0386

-0.128

0.3305

BYD - Zinc

0.1817

0.182

0.0125

0.1382

0.2113

BYD - Cobalt

0.0537

0.0519

0.0766

-0.4931

0.7751

BYD - Gold

0.0117

0.013

0.0193

-0.0664

0.0677

BYD - Silver

0.0633

0.0645

0.0344

-0.0972

0.2214

BYD - Platinum

0.0765

0.0795

0.0353

-0.1013

0.2298

BYD - Palladium

0.1331

0.1331

0.0203

-0.024

0.2001

CELS - Copper

0.1501

0.1459

0.0621

-0.0315

0.4234

CELS - Aluminium

0.0928

0.0922

0.0255

-0.0354

0.2293

CELS - Nickel

0.1047

0.1055

0.0585

-0.2156

0.463

CELS - Zinc

0.1196

0.1195

0.0158

0.0557

0.2077

CELS - Cobalt

0.0323

0.0331

0.0171

-0.0184

0.1061

CELS - Gold

0.0052

-0.0036

0.1432

-0.4706

0.4002

CELS - Silver

0.1117

0.1032

0.0845

-0.2035

0.3466

CELS - Platinum

0.1702

0.1659

0.1118

-0.0992

0.4215

CELS - Palladium

0.1832

0.1831

0.0012

0.1797

0.1879

SPGTCLEN - Copper

0.0063

0.0065

0.0172

-0.1042

0.1312

SPGTCLEN - Aluminium

-0.0027

-0.0027

0

-0.0027

-0.0027

SPGTCLEN - Nickel

0.0128

0.0133

0.0137

-0.0227

0.0476

SPGTCLEN - Zinc

-0.0305

-0.0274

0.011

-0.0698

-0.0046

SPGTCLEN - Cobalt

-0.0249

-0.0241

0.0253

-0.1542

0.0966

SPGTCLEN - Gold

-0.011

-0.0108

0.0415

-0.281

0.2464

SPGTCLEN - Silver

-0.0137

-0.0133

0.0575

-0.4207

0.319

SPGTCLEN - Platinum

0.0036

0.0034

0.0416

-0.2262

0.29

SPGTCLEN - Palladium

-0.0081

-0.0081

0

-0.0081

-0.0081

ECO - Copper

0.1542

0.1494

0.0761

-0.0676

0.4895

ECO - Aluminium

0.1012

0.1008

0.0416

-0.0654

0.2839

ECO - Nickel

0.1206

0.1204

0.056

-0.1156

0.4614

ECO - Zinc

0.1229

0.1228

0.0187

0.0569

0.229

ECO - Cobalt

0.0257

0.0259

0.0139

-0.0195

0.0941

ECO - Gold

0.0154

0.0089

0.1246

-0.3938

0.3517

ECO - Silver

0.127

0.121

0.0641

-0.0556

0.3088

ECO - Platinum

0.1787

0.1769

0.0995

-0.0767

0.4115

ECO - Palladium

0.1849

0.1849

0

0.1849

0.1849

Note: This table summarizes the mean, median, standard deviation, and range of time-varying correlations estimated using the
DCC-GARCH model for all equity–metal pairs. The results highlight differences in correlation strength and volatility across
EV manufacturers, traditional automakers, and clean energy indices.

80

Table 4 Diagnostic test
SERIES
CELS
SPGTCLEN
ECO
Tesla
BYD
Toyota
Honda
F
GM
Aluminium
Cobalt
Copper
Nickel
Zinc
Gold
Silver
Palladium
Platinum

ARCH LM (5)

ARCH LM (10)

LBQ (5)

LBQ (10)

543.5***

573.07***

13.03**

59.72***

0

0

-0.02

0

571.03***

593.01***

56.44***

86.13***

0

0

0

0

528.12***

559.34***

17.18***

57.87***

0

0

0

0

136.01***

147.47***

4.34

12.89

0

0

-0.5

-0.23

71.16***

77.92***

12.96**

17.83*

0

0

-0.02

-0.06

153.62***

208.25***

11.28**

18.18*

0

0

-0.05

-0.05

98.2***

137.51***

7.16

12.75

0

0

-0.21

-0.24

145.98***

206.76***

2.01

24.92***

0

0

-0.85

-0.01

491.26***

560.72***

4.41

45.74***

0

0

-0.49

0

122.47***

202.44***

24.98***

34.1***

0

0

0

0

47.09***

78.03***

40.96***

63.83***

0

0

0

0

208.6***

212.22***

6.79

17.27*

0

0

-0.24

-0.07

774.08***

781.04***

21.99***

29.87***

0

0

0

0

71.9***

84.79***

2.02

8.98

0

0

-0.85

-0.53

102.32***

138.97***

13.14**

22.54**

0

0

-0.02

-0.01

142.11***

156.52***

8.3

15.25

0

0

-0.14

-0.12

192.33***

462.93***

16.49***

49.49***

0

0

-0.01

0

301.19***

340***

9.15

33.49***

0

0

-0.1

0

Note: This table reports the results of the ARCH LM test and the Ljung–Box Q test for the return series under study. The
ARCH LM test (at lags 5 and 10) assesses the presence of conditional heteroscedasticity, while the Ljung–Box Q test examines
whether the model residuals exhibit autocorrelation. Statistically significant results (denoted by *, **, ***) suggest the
presence of ARCH effects or serial correlation, justifying the application of GARCH-type and multivariate time-series models
such as DCC-GARCH and VAR.

81

Overall, the diagnostic results support the application of multivariate time-series
models, particularly DCC-GARCH and VAR, to analyze the dynamic volatility structure and
cross-market dependencies among strategic metals, automakers, precious metals, and clean
energy indices.
4.4 Granger Causality Results
This section presents the results of Granger causality tests between strategic and
precious metals and various equity markets, including electric vehicle (EV) manufacturers,
traditional automakers, and clean energy indices. The goal is to assess whether past values of
metal returns can predict equity returns (Metal → Equity) and vice versa (Equity → Metal), as
pit market dynamics.
Table 5 presents the Granger causality test results for all pairs of strategic and precious
metals with equity markets, including EV manufacturers (Tesla, BYD), traditional automakers
(Toyota, Honda, GM, Ford), and clean energy indices (CELS, ECO, SPGTCLEN). The null
hypothesis tested is whether past returns of one asset fail to Granger-cause returns of the other.
A rejection of the null indicates statistically significant predictive power.
▪

EV Manufacturers (Tesla and BYD).

In the Metal → Equity direction, both Tesla and BYD exhibit statistically significant
sensitivity to multiple strategic and precious metal shocks. Tesla’s returns are Granger-caused
by Copper (p = 0.0916), Gold (p = 0.0358), Silver (p = 0.0009), and Platinum (p = 0.0081),
while BYD is significantly influenced by Cobalt, Copper, Nickel, Gold, Silver, Palladium, and

82

Platinum (p < 0.10). These findings indicate a strong predictive role of both battery-related and
safe-haven metals in EV valuation dynamics (Cagli, 2023; Erdoğan et al., 2022).
In the Equity → Metal direction, Tesla strongly Granger-causes the prices of Copper,
Nickel, Zinc, Aluminium, Palladium, and Platinum (p < 0.01), reflecting its status as a global
EV leader whose equity movements affect expectations in commodity markets (Kojić et al.,
2023). BYD also Granger-causes Copper and Zinc prices. This bidirectional causality
highlights the mutual interdependence between EV equity markets and strategic metal prices,
reinforcing the feedback dynamics crucial for risk forecasting and portfolio construction.
▪

Traditional Automakers (Toyota, Honda, Ford, GM):

Compared to EV firms, traditional automakers show weaker causality in the Metal →
Equity direction. Toyota is significantly affected by Zinc, Gold, and Palladium, and Honda by
Copper, Zinc, Gold, Palladium, and Platinum. In contrast, Ford and GM are influenced by a
broader set of metals, particularly Aluminium, Nickel, and multiple precious metals,
suggesting growing metal exposure as these firms expand EV production (Cagli, 2023;
Sovacool et al., 2020).
In the Equity → Metal direction, Ford and GM significantly Granger-cause Copper,
Nickel, Zinc, Palladium, and Platinum (p < 0.01), while Toyota and Honda show minimal
influence. These results confirm emerging bidirectional linkages for Ford and GM, consistent

83

with their strategic pivot toward electrification and deeper integration into metal supply chains
(Xing et al., 2023; Yahya et al., 2020).
▪

Clean Energy Indices (CELS, ECO, SPGTCLEN):

CELS and ECO show strong Metal → Equity causality from Aluminium, Copper, and
Nickel (p < 0.01), as well as from multiple precious metals. These results highlight their
dependence on metal-intensive technologies such as wind turbines, solar panels, and storage
systems (IEA, 2021; Sovacool et al., 2020). In contrast, SPGTCLEN displays limited
sensitivity, with significant causality only from Aluminium, Copper, and Platinum.
In the reverse direction, CELS and ECO significantly Granger-cause the prices of key
metals including Aluminium, Copper, Nickel, and Palladium. The strength of these
bidirectional links suggests that clean energy equity returns not only reflect but also shape
expectations in metals markets through anticipated demand shifts(Erdoğan et al., 2022; Yahya
et al., 2020). SPGTCLEN shows no significant reverse causality, possibly due to broader
diversification and weaker ties to specific clean energy inputs.
Fig. 6 presents a directional matrix summarizing the statistically significant Granger
causal relationships between strategic and precious metals and various equity markets, as tested
in Table 5. The Granger causality analysis of strategic (non-precious) metals reveals
heterogeneous and firm-specific causal directions. Copper shows bidirectional causality with
both Tesla and BYD, suggesting strong two-way feedback between EV stock dynamics and
copper price movements. Aluminium displays no significant causality with either firm,
indicating limited strategic influence in short-run dynamics. Nickel shows no significant link
with Tesla, but does Granger-cause BYD, implying that BYD is more exposed to nickel
84

volatility, likely due to differences in battery technology or sourcing strategies (Lander et al.,
2021). Zinc reveals a one-way effect from the stock returns of Toyota and Honda to zinc prices,
whereas a bidirectional relationship exists with the CELS index, highlighting the mutual
influence between energy transition stocks and industrial metal prices. Notably, Cobalt
Granger-causes BYD (Metal → Equity), while GM and Honda appear to influence cobalt
prices instead (Equity → Metal), reinforcing the asymmetric nature of commodity-equity
interaction.
Table 5 Granger Causality Test Results Between Strategic/Precious Metals and Equity Markets (Bidirectional
Analysis)
Pair

Null Hypothesis

Lag

BYD - Aluminium

Aluminium → BYD

3

BYD - Aluminium

BYD → Aluminium

3

BYD - Cobalt

Cobalt → BYD

5

BYD - Cobalt

BYD → Cobalt

5

BYD - Copper

Copper → BYD

1

BYD - Copper

BYD → Copper

1

BYD - Nickel

Nickel → BYD

3

BYD - Nickel

BYD → Nickel

3

BYD - Zinc

Zinc → BYD

3

BYD - Zinc

BYD → Zinc

3

BYD - Gold

Gold → BYD

3

BYD - Gold

BYD → Gold

3

BYD - Silver

Silver → BYD

3

BYD - Silver

BYD → Silver

3

BYD - Palladium

Palladium → BYD

3

BYD - Palladium

BYD → Palladium

3

BYD - Platinum

Platinum → BYD

1

BYD - Platinum

BYD → Platinum

1

Tesla - Aluminium

Aluminium → Tesla

3

85

F Statistic
1.787
(0.1474)
2.121*
(0.0953)
1.886*
(0.0933)
0.886
(0.4895)
2.808*
(0.0939)
5.319**
(0.0211)
2.265*
(0.0789)
1.115
(0.3416)
0.685
(0.5610)
2.735**
(0.0421)
3.043**
(0.0277)
0.507
(0.6774)
5.351***
(0.0011)
0.558
(0.6428)
4.525***
(0.0036)
1.236
(0.2950)
15.181***
(0.0001)
0.442
(0.5063)
1.438
(0.2297)

Caused By

Decision

Aluminium

Fail to Reject

BYD

Reject H₀

Cobalt

Reject H₀

BYD

Fail to Reject

Copper

Reject H₀

BYD

Reject H₀

Nickel

Reject H₀

BYD

Fail to Reject

Zinc

Fail to Reject

BYD

Reject H₀

Gold

Reject H₀

BYD

Fail to Reject

Silver

Reject H₀

BYD

Fail to Reject

Palladium

Reject H₀

BYD

Fail to Reject

Platinum

Reject H₀

BYD

Fail to Reject

Aluminium

Fail to Reject

Tesla - Aluminium

Tesla → Aluminium

3

Tesla - Cobalt

Cobalt → Tesla

2

Tesla - Cobalt

Tesla → Cobalt

2

Tesla - Copper

Copper → Tesla

1

Tesla - Copper

Tesla → Copper

1

Tesla - Nickel

Nickel → Tesla

3

Tesla - Nickel

Tesla → Nickel

3

Tesla - Zinc

Zinc → Tesla

1

Tesla - Zinc

Tesla → Zinc

1

Tesla - Gold

Gold → Tesla

4

Tesla - Gold

Tesla → Gold

4

Tesla - Silver

Silver → Tesla

4

Tesla - Silver

Tesla → Silver

4

Tesla - Palladium

Palladium → Tesla

1

Tesla - Palladium

Tesla → Palladium

1

Tesla - Platinum

Platinum → Tesla

2

Tesla - Platinum

Tesla → Platinum

2

Toyota - Aluminium

Aluminium → Toyota

2

Toyota - Aluminium

Toyota → Aluminium

2

Toyota - Cobalt

Cobalt → Toyota

2

Toyota - Cobalt

Toyota → Cobalt

2

Toyota - Copper

Copper → Toyota

1

9.544***
(0.0000)
0.299
(0.7415)
0.578
(0.5610)
2.846*
(0.0916)
43.36***
(0.0000)
1.933
(0.1219)
11.717***
(0.0000)
0.47 (0.4928)
32.158***
(0.0000)
2.575**
(0.0358)
0.909
(0.4577)
4.704***
(0.0009)
0.573
(0.6820)
0.045
(0.8328)
9.563***
(0.0020)
4.815***
(0.0081)
3.912**
(0.0201)
0.397
(0.6725)
0.458
(0.6327)
0.266
(0.7666)
1.356
(0.2578)
1.65 (0.1990)

Toyota - Copper

Toyota → Copper

1

Toyota - Nickel

Nickel → Toyota

1

0.457
(0.4989)
0 (0.9938)

1

0.01 (0.9206)

Toyota - Nickel

Toyota → Nickel

Toyota - Zinc

Zinc → Toyota

1

Toyota - Zinc

Toyota → Zinc

1

Toyota - Gold

Gold → Toyota

3

Toyota - Gold

Toyota → Gold

3

Toyota - Silver

Silver → Toyota

1

Toyota - Silver

Toyota → Silver

1

Toyota - Palladium

Palladium → Toyota

1

86

4.708**
(0.0301)
0.081
(0.7756)
6.972***
(0.0001)
0.147
(0.9314)
1.644
(0.1998)
0.612
(0.4342)
10.699***
(0.0011)

Tesla

Reject H₀

Cobalt

Fail to Reject

Tesla

Fail to Reject

Copper

Reject H₀

Tesla

Reject H₀

Nickel

Fail to Reject

Tesla

Reject H₀

Zinc

Fail to Reject

Tesla

Reject H₀

Gold

Reject H₀

Tesla

Fail to Reject

Silver

Reject H₀

Tesla

Fail to Reject

Palladium

Fail to Reject

Tesla

Reject H₀

Platinum

Reject H₀

Tesla

Reject H₀

Aluminium

Fail to Reject

Toyota

Fail to Reject

Cobalt

Fail to Reject

Toyota

Fail to Reject

Copper

Fail to Reject

Toyota

Fail to Reject

Nickel

Fail to Reject

Toyota

Fail to Reject

Zinc

Reject H₀

Toyota

Fail to Reject

Gold

Reject H₀

Toyota

Fail to Reject

Silver

Fail to Reject

Toyota

Fail to Reject

Palladium

Reject H₀

Toyota - Palladium

Toyota → Palladium

1

Toyota - Platinum

Platinum → Toyota

1

Toyota - Platinum

Toyota → Platinum

1

Honda - Aluminium

Aluminium → Honda

3

Honda - Aluminium

Honda → Aluminium

3

Honda - Cobalt

Cobalt → Honda

2

Honda - Cobalt

Honda → Cobalt

2

Honda - Copper

Copper → Honda

1

Honda - Copper

Honda → Copper

1

Honda - Nickel

Nickel → Honda

1

Honda - Nickel

Honda → Nickel

1

Honda - Zinc

Zinc → Honda

1

Honda - Zinc

Honda → Zinc

1

Honda - Gold

Gold → Honda

3

Honda - Gold

Honda → Gold

3

Honda - Silver

Silver → Honda

4

Honda - Silver

Honda → Silver

4

Honda - Palladium

Palladium → Honda

1

Honda - Palladium

Honda → Palladium

1

Honda - Platinum

Platinum → Honda

1

Honda - Platinum

Honda → Platinum

1

F - Aluminium

Aluminium → F

3

F - Aluminium

F → Aluminium

3

F - Cobalt

Cobalt → F

2

F - Cobalt

F → Cobalt

2

F - Copper

Copper → F

1

F - Copper

F → Copper

1

F - Nickel

Nickel → F

3

F - Nickel

F → Nickel

3

F - Zinc

Zinc → F

1

F - Zinc

F → Zinc

1

87

3.927**
(0.0476)
1.787
(0.1813)
0.332
(0.5643)
1.254
(0.2883)
0.023
(0.9954)
0.277
(0.7582)
2.422*
(0.0889)
8.911***
(0.0028)
0.211
(0.6463)
0.116
(0.7332)
0.951
(0.3294)
4.67**
(0.0307)
1.101
(0.2941)
6.154***
(0.0004)
0.59 (0.6213)
1.362
(0.2446)
1.589
(0.1742)
14.329***
(0.0002)
1.515
(0.2185)
5.519**
(0.0189)
2.484
(0.1150)
2.94**
(0.0319)
14.704***
(0.0000)
0.441
(0.6433)
2.209
(0.1099)
2.432
(0.1189)
37.776***
(0.0000)
2.454*
(0.0614)
13.917***
(0.0000)
2.007
(0.1567)
43.142***
(0.0000)

Toyota

Reject H₀

Platinum

Fail to Reject

Toyota

Fail to Reject

Aluminium

Fail to Reject

Honda

Fail to Reject

Cobalt

Fail to Reject

Honda

Reject H₀

Copper

Reject H₀

Honda

Fail to Reject

Nickel

Fail to Reject

Honda

Fail to Reject

Zinc

Reject H₀

Honda

Fail to Reject

Gold

Reject H₀

Honda

Fail to Reject

Silver

Fail to Reject

Honda

Fail to Reject

Palladium

Reject H₀

Honda

Fail to Reject

Platinum

Reject H₀

Honda

Fail to Reject

Aluminium

Reject H₀

F

Reject H₀

Cobalt

Fail to Reject

F

Fail to Reject

Copper

Fail to Reject

F

Reject H₀

Nickel

Reject H₀

F

Reject H₀

Zinc

Fail to Reject

F

Reject H₀

F - Gold

Gold → F

2

F - Gold

F → Gold

2

F - Silver

Silver → F

2

F - Silver

F → Silver

2

F - Palladium

Palladium → F

1

F - Palladium

F → Palladium

1

F - Platinum

Platinum → F

2

F - Platinum

F → Platinum

2

GM - Aluminium

Aluminium → GM

3

GM - Aluminium

GM → Aluminium

3

GM - Cobalt

Cobalt → GM

2

GM - Cobalt

GM → Cobalt

2

GM - Copper

Copper → GM

1

GM - Copper

GM → Copper

1

GM - Nickel

Nickel → GM

3

GM - Nickel

GM → Nickel

3

GM - Zinc

Zinc → GM

1

GM - Zinc

GM → Zinc

1

GM - Gold

Gold → GM

3

GM - Gold

GM → Gold

3

GM - Silver

Silver → GM

2

GM - Silver

GM → Silver

2

GM - Palladium

Palladium → GM

1

GM - Palladium

GM → Palladium

1

GM - Platinum

Platinum → GM

2

GM - Platinum

GM → Platinum

2

CELS - Aluminium

Aluminium → CELS

2

CELS - Aluminium

CELS → Aluminium

2

CELS - Cobalt

Cobalt → CELS

2

CELS - Cobalt

CELS → Cobalt

2

CELS - Copper

Copper → CELS

2

88

5.816***
(0.0030)
2.214
(0.1094)
7.844***
(0.0004)
2.178
(0.1133)
0.933
(0.3341)
23.596***
(0.0000)
3.831**
(0.0217)
6.987***
(0.0009)
2.553*
(0.0538)
12.633***
(0.0000)
0.395
(0.6736)
4.818***
(0.0081)
2.345
(0.1257)
42.008***
(0.0000)
3.479**
(0.0153)
11.193***
(0.0000)
2.55 (0.1103)
27.029***
(0.0000)
2.615**
(0.0494)
3.25**
(0.0209)
8.024***
(0.0003)
4.961***
(0.0070)
3.372*
(0.0664)
27.813***
(0.0000)
3.803**
(0.0224)
8.428***
(0.0002)
6.532***
(0.0015)
36.494***
(0.0000)
0.079
(0.9238)
1.386
(0.2501)
8.896***
(0.0001)

Gold

Reject H₀

F

Fail to Reject

Silver

Reject H₀

F

Fail to Reject

Palladium

Fail to Reject

F

Reject H₀

Platinum

Reject H₀

F

Reject H₀

Aluminium

Reject H₀

GM

Reject H₀

Cobalt

Fail to Reject

GM

Reject H₀

Copper

Fail to Reject

GM

Reject H₀

Nickel

Reject H₀

GM

Reject H₀

Zinc

Fail to Reject

GM

Reject H₀

Gold

Reject H₀

GM

Reject H₀

Silver

Reject H₀

GM

Reject H₀

Palladium

Reject H₀

GM

Reject H₀

Platinum

Reject H₀

GM

Reject H₀

Aluminium

Reject H₀

CELS

Reject H₀

Cobalt

Fail to Reject

CELS

Fail to Reject

Copper

Reject H₀

CELS - Copper

CELS → Copper

2

CELS - Nickel

Nickel → CELS

3

CELS - Nickel

CELS → Nickel

3

CELS - Zinc

Zinc → CELS

2

CELS - Zinc

CELS → Zinc

2

CELS - Gold

Gold → CELS

2

CELS - Gold

CELS → Gold

2

CELS - Silver

Silver → CELS

2

CELS - Silver

CELS → Silver

2

CELS - Palladium

Palladium → CELS

2

CELS - Palladium

CELS → Palladium

2

CELS - Platinum

Platinum → CELS

2

CELS - Platinum

CELS → Platinum

2

ECO - Aluminium

Aluminium → ECO

2

ECO - Aluminium

ECO → Aluminium

2

ECO - Cobalt

Cobalt → ECO

2

ECO - Cobalt

ECO → Cobalt

2

ECO - Copper

Copper → ECO

2

ECO - Copper

ECO → Copper

2

ECO - Nickel

Nickel → ECO

3

ECO - Nickel

ECO → Nickel

3

ECO - Zinc

Zinc → ECO

2

ECO - Zinc

ECO → Zinc

2

ECO - Gold

Gold → ECO

2

ECO - Gold

ECO → Gold

2

ECO - Silver

Silver → ECO

2

ECO - Silver

ECO → Silver

2

ECO - Palladium

Palladium → ECO

2

ECO - Palladium

ECO → Palladium

2

ECO - Platinum

Platinum → ECO

2

ECO - Platinum

ECO → Platinum

2

89

37.615***
(0.0000)
4.571***
(0.0033)
15.284***
(0.0000)
3.077**
(0.0462)
31.615***
(0.0000)
3.133**
(0.0437)
1.648
(0.1924)
3.939**
(0.0195)
5.871***
(0.0028)
2.451*
(0.0863)
27.298***
(0.0000)
4.162**
(0.0156)
11.688***
(0.0000)
4.752***
(0.0087)
33.942***
(0.0000)
0.081
(0.9221)
1.097
(0.3339)
7.072***
(0.0009)
33.131***
(0.0000)
3.063**
(0.0270)
11.01***
(0.0000)
2.121
(0.1200)
28.511***
(0.0000)
2.584*
(0.0756)
2.483*
(0.0836)
2.69*
(0.0680)
5.732***
(0.0033)
3.193**
(0.0411)
28.368***
(0.0000)
2.486*
(0.0834)
11.49***
(0.0000)

CELS

Reject H₀

Nickel

Reject H₀

CELS

Reject H₀

Zinc

Reject H₀

CELS

Reject H₀

Gold

Reject H₀

CELS

Fail to Reject

Silver

Reject H₀

CELS

Reject H₀

Palladium

Reject H₀

CELS

Reject H₀

Platinum

Reject H₀

CELS

Reject H₀

Aluminium

Reject H₀

ECO

Reject H₀

Cobalt

Fail to Reject

ECO

Fail to Reject

Copper

Reject H₀

ECO

Reject H₀

Nickel

Reject H₀

ECO

Reject H₀

Zinc

Fail to Reject

ECO

Reject H₀

Gold

Reject H₀

ECO

Reject H₀

Silver

Reject H₀

ECO

Reject H₀

Palladium

Reject H₀

ECO

Reject H₀

Platinum

Reject H₀

ECO

Reject H₀

SPGTCLEN - Aluminium

Aluminium → SPGTCLEN

2

SPGTCLEN - Aluminium

SPGTCLEN → Aluminium

2

SPGTCLEN - Cobalt

Cobalt → SPGTCLEN

5

SPGTCLEN - Cobalt

SPGTCLEN → Cobalt

5

SPGTCLEN - Copper

Copper → SPGTCLEN

2

SPGTCLEN - Copper

SPGTCLEN → Copper

2

SPGTCLEN - Nickel

Nickel → SPGTCLEN

2

SPGTCLEN - Nickel

SPGTCLEN → Nickel

2

SPGTCLEN - Zinc

Zinc → SPGTCLEN

2

SPGTCLEN - Zinc

SPGTCLEN → Zinc

2

SPGTCLEN - Gold

Gold → SPGTCLEN

3

SPGTCLEN - Gold

SPGTCLEN → Gold

3

SPGTCLEN - Silver

Silver → SPGTCLEN

4

4.752***
(0.0087)
0.888
(0.4114)
1.253
(0.2817)
0.565
(0.7268)
2.513*
(0.0812)
0.762
(0.4670)
1.14 (0.3199)
0.154
(0.8577)
1.512
(0.2206)
0.627
(0.5341)
1.441
(0.2288)
0.357
(0.7840)
1.27 (0.2793)

Aluminium

Reject H₀

SPGTCLEN

Fail to Reject

Cobalt

Fail to Reject

SPGTCLEN

Fail to Reject

Copper

Reject H₀

SPGTCLEN

Fail to Reject

Nickel

Fail to Reject

SPGTCLEN

Fail to Reject

Zinc

Fail to Reject

SPGTCLEN

Fail to Reject

Gold

Fail to Reject

SPGTCLEN

Fail to Reject

Silver
Fail to Reject
0.449
SPGTCLEN - Silver
SPGTCLEN → Silver
4
SPGTCLEN
Fail to Reject
(0.7728)
0.628
SPGTCLEN - Palladium
Palladium → SPGTCLEN
2
Palladium
Fail to Reject
(0.5335)
0.343
SPGTCLEN - Palladium
SPGTCLEN → Palladium
2
SPGTCLEN
Fail to Reject
(0.7098)
3.601**
SPGTCLEN - Platinum
Platinum → SPGTCLEN
2
Platinum
Reject H₀
(0.0274)
0.706
SPGTCLEN - Platinum
SPGTCLEN → Platinum
2
SPGTCLEN
Fail to Reject
(0.4938)
Note: This table summarizes the results of Granger causality tests assessing the bidirectional predictive relationships between
strategic and precious metals and equity markets, including electric vehicle (EV) manufacturers, traditional automakers, and
clean energy indices. The direction "Metal → Equity" tests whether metal price returns help predict equity returns, while
"Equity → Metal" tests the reverse. The F-statistics show the test value, and the p-value (in parentheses) indicates statistical
significance. Asterisks denote significance levels: *** p < 0.01, ** p < 0.05, * p < 0.10. The null hypothesis H0H_0H0 assumes
no Granger-causal relationship.

In contrast, the Granger causality relationships involving precious metals exhibit more
consistent and passive dynamics, aligning with their traditional role as safe-haven assets. Gold
Granger-causes both Tesla and BYD, suggesting that during periods of uncertainty, movements
in gold prices influence EV stock performance, while the reverse is not observed. Silver shows
bidirectional relationships with CELS, ECO, and GM, and one-way causality (Metal →
Equity) with Tesla, BYD, and Ford, indicating that silver prices can predict stock movements,
especially due to its industrial use in clean technologies like solar energy (Dutta, 2019; Lahiani
et al., 2021). Platinum also Granger-causes both EV firms, reflecting its influence as a stable
90

hedging instrument with moderate volatility. In contrast, Palladium exhibits stock-to-metal
causality with Tesla, and unidirectional causality (Metal → Equity) with BYD, suggesting that
while Tesla may influence palladium markets through demand signals, BYD is more reactive
to changes in palladium prices.

Fig. 6 Granger Causality Directions Between Metals and Equities
Note: This figure illustrates the statistically significant Granger Causality relationships between strategic and precious metals
and equity markets, based on results from Table 5. Blue arrows (→) indicate that metal returns Granger-cause equity returns
(Metal → Equity), red arrows (←) indicate equity returns Granger-cause metal returns (Equity → Metal), and purple
bidirectional arrows (↔) represent significant causality in both directions. Grey cells denote no statistically significant
causality at the 10% level. These results are based on pairwise VAR models with optimal lag selection via AIC and Granger
causality tests at 5% and 10% significance thresholds over the period January 2014–January 2024.

4.5 Impulse Response Functions (IRF) Results
This section presents the Impulse Response Function (IRF) analysis based on the VAR
models estimated in the previous section. The IRF traces the dynamic response of equity
markets—including electric vehicle (EV) manufacturers, traditional automakers, and clean
energy indices—to one-standard-deviation shocks in strategic and precious metal returns.
91

These results help assess the direction, magnitude, and persistence of metal price shocks on
equity returns over a 10-day horizon, offering deeper insights beyond Granger causality by
illustrating the temporal transmission of market effects(Lütkepohl, 2005).
Fig. 7 presents the impulse response functions (IRFs) capturing the dynamic effects of
one-standard-deviation shocks in strategic (non-precious) and precious metal prices on equity
returns across electric vehicle (EV) manufacturers, traditional automakers, and clean energy
indices. Each panel illustrates how a specific equity responds to a metal price shock over a 10day horizon, with 95% confidence intervals (dashed lines). These IRFs complement the
Granger causality results presented in Table 5 and the directional matrix in Fig. 6 by providing
temporal insights into the magnitude and persistence of market reactions.
For example, a shock in Copper prices cause a statistically significant and short-lived
increase in Tesla’s returns, peaking on day 1 and fading by day 2. In contrast, Aluminium
shocks produce a significant but brief response in Ford’s equity returns, with effects dissipating
by period 5. These examples highlight how strategic metals affect different equity groups in
distinct ways, lending dynamic support to Hypothesis H3 concerning the sensitivity of EV and
clean energy equities to metal price volatility.
Additionally, IRFs for precious metals reveal distinct dynamics. For instance, a gold
price shock causes a significant negative response in Tesla’s returns, reaching its lowest point
around period 4 before gradually reverting to zero, highlighting Tesla’s potential exposure to
gold-driven risk sentiment. Conversely, GM reacts positively to silver price shocks, with a
significant rise in returns sustained until approximately period 5, after which the effect
diminishes. These patterns provide complementary insights into the hedging and sentimentrelated behavior of precious metals in equity markets.
92

93

94

95

96

Fig. 7 Impulse Response Functions (IRFs): Metal Shocks to Equity Reactions
Note: This figure illustrates how equity returns respond over a 10-day period following a one-standard-deviation shock in
strategic and precious metal prices. The results show firm- and sector-specific reactions: for example, Tesla reacts immediately
to copper shocks, while GM responds positively and persistently to silver. These patterns highlight the time-varying and
asymmetrical influence of metal markets on equity performance.

4.6 Hedging Analysis and Portfolio Weights
This section presents the results of the hedging analysis and optimal portfolio weight
estimation between strategic and precious metals and various equity sectors, including electric
vehicle (EV) manufacturers, traditional automakers, and clean energy indices. Based on the
DCC-GARCH model, these findings respond directly to Hypotheses H4 and H5, which posit
(i) stronger hedge linkages for strategic metals and (ii) the cost-effectiveness of precious metals
in hedging under crisis conditions.

97

The hedge ratio (β) quantifies the proportion of a metal position required to hedge against
equity market exposure and is calculated using the conditional covariance and variance from
the DCC-GARCH estimates:
𝐶𝑜𝑣𝑎𝑏,𝑡
(31)
𝑉𝑎𝑟𝑏𝑏,𝑡
Where Covab,t is the conditional covariance between asset a (metal) and asset b (equity
𝛽𝑎𝑏,𝑡 =

index), and Varbb,t is the conditional variance of the equity index at time t. A higher hedge
ratio implies a stronger hedging demand and greater co-movement between the asset and the
hedge instrument.
The conditional variance and covariances calculated from the DCC-GARCH model
guide the construction of optimal portfolio weights, showing the best mix of clean energy
subsectors and precious metal commodities for investors optimizing their portfolios. Following
the approaches by Arouri et al., (2011) and Kroner and Ng (1998), the optimal weights for
holding two assets are:

𝑊𝑎𝑏,𝑡 =

𝑉𝑎𝑟𝑏𝑏,𝑡 − 𝐶𝑜𝑣𝑎𝑏,𝑡
𝑉𝑎𝑟𝑎𝑎,𝑡 − 2 𝐶𝑜𝑣𝑎𝑏,𝑡 + 𝑉𝑎𝑟𝑏𝑏,𝑡

(32)

𝑖𝑓 𝑊𝑎𝑏,𝑡 < 0,
(33)
𝑖𝑓 0 ≤ 𝑊𝑎𝑏,𝑡 ≤ 1,
𝑊𝑎𝑏,𝑡 ={ 𝑊𝑎𝑏,𝑡
1
𝑖𝑓 𝑊𝑎𝑏,𝑡 > 1,
where Wab,t , indicates the weight of the first asset (e.g., equity index) in 1$ portfolio of
0

two assets (e.g., equity index and metal commodity) at time t. 1 − Wab,t , can be used to calculate
the weight of the second asset in the portfolio.
The dynamic hedging relationships between strategic and precious metals and various
equity markets, electric vehicle (EV) manufacturers, traditional automakers, and clean energy
98

indices, are illustrated in Fig. 8 and summarized in Table 6, offering key insights into hedge
cost dynamics and market interconnectedness over the 2014–2024 period. EV stocks such as
Tesla and BYD show significantly different hedging profiles. Gold consistently exhibits the
lowest average hedge ratios (Tesla–Gold = 0.06; BYD–Gold = 0.04), indicating that a small
short position in gold is sufficient to hedge a $1 long position in EV equities. This supports
gold’s well-documented function as a passive hedge and safe haven asset, due to its weak
correlation with equity volatility (Baur & Lucey, 2010b).
Conversely, strategic metals like Cobalt and Copper show significantly higher average
hedge ratios for BYD (Cobalt = 0.45; Copper = 0.50) than for Tesla (0.32 and 0.22
respectively), indicating more expensive hedging requirements and deeper financial exposure.
This discrepancy likely reflects BYD’s vertically integrated battery supply chain, which
amplifies its vulnerability to raw material price shocks (Pham & Hsu, 2025; Reboredo, 2013;
Sari et al., 2010). The COVID-19 pandemic and Russia–Ukraine war periods show visible
spikes in hedge ratios (Fig. 8), particularly in Tesla–Cobalt and BYD–Copper pairs,
underscoring the sensitivity of EV-metal linkages to global disruptions.
Fig. 9 further reveals that strategic metals display higher median hedge ratios and wider
interquartile ranges during crisis periods, particularly in EV pairings with Cobalt and Copper.
In contrast, precious metals such as Gold and Silver maintain narrow, stable distributions with
low medians, reinforcing their role as low-cost, passive hedging instruments across regimes.
Among traditional automakers, strategic metals such as Copper, Aluminium, and Zinc
consistently exhibit higher average hedge ratios than precious metals such as Gold and Silver.
For instance, the GM–Copper pair shows the highest mean hedge ratio (0.25), followed by
Ford–Copper (0.20) and Honda–Copper (0.17), indicating stronger co-movement and higher
99

100

101

Fig. 8 Time‐varying hedge ratios between equity indices and metals estimated via DCC-GARCH models.

102

Note: This figure illustrates the evolution of optimal hedge ratios from 2014 to 2024, capturing the dynamic risk relationships
between metals and equities. Spikes in hedge ratios during COVID-19 and the Russia–Ukraine war reflect increased comovements and hedging costs under market stress.

hedging demand. Gold, on the other hand, shows consistently negative hedge ratios (Toyota =
–0.13; Honda = –0.14; GM = –0.12), affirming its inverse relationship with equity returns and
usefulness as a safe haven (Baur & Lucey, 2010b).Fig. 9 supports this pattern, showing that
during the COVID-19 crisis, hedge ratios for GM–Nickel and GM–Zinc surged, reflecting
increased commodity sensitivity during global shocks (Alekseev et al., 2024).
Clean energy indices display sector-specific hedging behavior. As per Table 6, ECO
and CELS exhibit higher average hedge ratios with strategic metals (ECO–Copper = 0.26;
CELS–Cobalt = 0.21; CELS–Copper = 0.23), indicating elevated exposure to commodity price
risks due to technological dependence on metals for solar, wind, and battery infrastructure
(Gustafsson et al., 2022). In contrast, the globally diversified SPGTCLEN index shows
negligible or negative ratios with many metals (e.g., SPGTCLEN–Cobalt = –0.11), suggesting
lower vulnerability. As Fig. 9 shows, clean energy indices experience elevated hedge ratios
Table 6 summary statistics for the optimal hedge ratios across different equity indices and metals
Pair

Mean

Median

SD

Min

Max

Toyota/Copper

0.13

0.12

0.05

-0.02

0.50

Toyota/Aluminium

0.10

0.09

0.05

0.02

0.42

Toyota/Nickel

0.07

0.06

0.06

-0.21

0.56

Toyota/Zinc

0.10

0.09

0.03

0.05

0.29

Toyota/Cobalt

0.16

0.08

0.20

-0.23

2.11

Toyota/Gold

-0.13

-0.12

0.08

-1.23

0.23

Toyota/Silver

-0.01

-0.01

0.01

-0.10

0.03

Toyota/Platinum

0.01

0.01

0.03

-0.26

0.35

Toyota/Palladium

0.04

0.04

0.03

-0.05

0.35

Honda/Copper

0.17

0.16

0.08

0.03

0.54

Honda/Aluminium

0.11

0.10

0.03

0.04

0.26

Honda/Nickel

0.09

0.08

0.04

-0.04

0.27

Honda/Zinc

0.11

0.11

0.03

0.05

0.22

Honda/Cobalt

0.05

0.03

0.07

-0.06

0.62

Honda/Gold

-0.14

-0.13

0.07

-0.62

0.20

103

Honda/Silver

-0.01

-0.01

0.03

-0.13

0.20

Honda/Platinum

0.02

0.02

0.05

-0.27

0.37

Honda/Palladium

0.05

0.05

0.03

-0.07

0.25

GM/Copper

0.25

0.21

0.20

-0.08

1.67

GM/Aluminium

0.11

0.09

0.12

-0.05

1.43

GM/Nickel

0.10

0.07

0.17

-0.13

1.71

GM/Zinc

0.11

0.09

0.14

-0.10

1.45

GM/Cobalt

-0.02

-0.01

0.05

-0.33

0.40

GM/Gold

-0.12

-0.11

0.21

-1.06

0.74

GM/Silver

0.07

0.06

0.08

-0.20

0.53

GM/Platinum

0.19

0.19

0.15

-0.22

0.83

GM/Palladium

0.14

0.14

0.03

0.07

0.27

F/Copper

0.20

0.18

0.11

-0.16

1.13

F/Aluminium

0.11

0.11

0.05

0.03

0.46

F/Nickel

0.08

0.07

0.08

-0.12

0.98

F/Zinc

0.12

0.10

0.08

-0.08

0.82

F/Cobalt

0.02

0.01

0.04

-0.17

0.35

F/Gold

-0.11

-0.11

0.20

-0.81

0.68

F/Silver

0.06

0.05

0.08

-0.21

0.33

F/Platinum

0.17

0.17

0.12

-0.14

0.54

F/Palladium

0.13

0.12

0.04

0.03

0.31

Tesla/Copper

0.22

0.21

0.08

0.00

0.78

Tesla/Aluminium

0.10

0.09

0.10

-0.41

1.03

Tesla/Nickel

0.13

0.11

0.08

-0.09

0.92

Tesla/Zinc

0.17

0.16

0.09

-0.29

1.05

Tesla/Cobalt

0.32

0.13

0.57

-0.43

5.74

Tesla/Gold

0.06

0.05

0.11

-0.58

0.71

Tesla/Silver

0.13

0.13

0.05

-0.16

0.61

Tesla/Platinum

0.27

0.25

0.08

0.09

0.66

Tesla/Palladium

0.21

0.20

0.06

0.08

0.47

BYD/Copper

0.50

0.48

0.26

-0.26

1.66

BYD/Aluminium

0.35

0.31

0.13

0.12

0.88

BYD/Nickel

0.28

0.25

0.14

-0.10

1.10

BYD/Zinc

0.37

0.33

0.14

0.14

1.03

BYD/Cobalt

0.45

0.24

0.52

-1.34

2.92

BYD/Gold

0.04

0.04

0.07

-0.28

0.29

BYD/Silver

0.11

0.10

0.07

-0.21

0.45

BYD/Platinum

0.15

0.14

0.10

-0.43

0.79

BYD/Palladium

0.20

0.18

0.09

-0.03

0.53

CELS/Copper

0.23

0.21

0.17

-0.07

1.60

CELS/Aluminium

0.14

0.12

0.11

-0.08

1.42

CELS/Nickel

0.11

0.08

0.11

-0.15

1.93

CELS/Zinc

0.15

0.13

0.08

0.05

0.95

CELS/Cobalt

0.21

0.07

0.41

-0.18

4.62

104

CELS/Gold

0.03

-0.01

0.29

-0.88

1.20

CELS/Silver

0.12

0.10

0.10

-0.11

0.64

CELS/Platinum

0.21

0.19

0.15

-0.13

0.86

CELS/Palladium

0.16

0.15

0.05

0.06

0.49

SPGTCLEN/Copper

0.01

0.01

0.02

-0.20

0.35

SPGTCLEN/Aluminium

0.00

0.00

0.00

-0.02

0.00

SPGTCLEN/Nickel

0.01

0.01

0.02

-0.02

0.20

SPGTCLEN/Zinc

-0.03

-0.02

0.01

-0.09

0.00

SPGTCLEN/Cobalt

-0.11

-0.04

0.17

-0.89

0.30

SPGTCLEN/Gold

-0.02

-0.01

0.07

-0.55

0.43

SPGTCLEN/Silver

-0.01

-0.01

0.05

-0.59

0.23

SPGTCLEN/Platinum

0.00

0.00

0.04

-0.31

0.36

SPGTCLEN/Palladium

-0.01

0.00

0.00

-0.02

0.00

ECO/Copper

0.26

0.23

0.20

-0.15

1.92

ECO/Aluminium

0.17

0.15

0.15

-0.20

1.71

ECO/Nickel

0.13

0.11

0.13

-0.09

1.92

ECO/Zinc

0.16

0.15

0.09

0.05

1.01

ECO/Cobalt

0.19

0.06

0.37

-0.22

4.10

ECO/Gold

0.06

0.02

0.29

-0.76

1.15

ECO/Silver

0.14

0.12

0.09

-0.06

0.53

ECO/Platinum

0.24

0.22

0.15

-0.09

0.80

ECO/Palladium

0.18

0.17

0.06

0.07

0.48

Note: This table summarizes the mean, median, standard deviation, minimum, and maximum of optimal hedge ratios between
equities and metals from 2014–2024. Higher hedge ratios imply costlier hedging, while negative values reflect potential
inverse hedging relationships.

with strategic metals during COVID-19, particularly in ECO–Cobalt and CELS–Copper,
reinforcing their crisis-responsiveness. Meanwhile, precious metals such as Gold and Silver
maintain consistently low hedge ratios (e.g., CELS–Gold = 0.03, ECO–Gold = 0.06),
highlighting their role as stable, low-cost hedges across market states.
Overall, these findings validate the thesis hypotheses that:
1) Strategic metals exhibit higher hedge costs and co-movement, particularly during
crises.
2) EV and clean energy stocks are more exposed to metal price risks than traditional
automakers.
3) Precious metals offer cost-effective hedging, particularly in systemic crises.

105

These insights suggest that dynamic and asset-specific hedging strategies are essential
for portfolio managers in sectors exposed to commodity risk, especially amid geopolitical
shocks and supply chain disruptions.
Table 7 reports the summary statistics of optimal portfolio weights between equity
indices, including electric vehicle (EV) manufacturers (Tesla, BYD), traditional automakers
(Toyota, Honda, GM, Ford), and clean energy indices (CELS, ECO, SPGTCLEN), and both
strategic and precious metals over the 2014–2024 period. The optimal weights, derived from
DCC-GARCH conditional variances and covariances, indicate the proportion of equity
allocated within a two-asset portfolio, with the remaining weight implicitly assigned to the
hedging metal. These values offer key insights into the co-movement dynamics and hedging
contributions of each metal across different sectors.
Within the EV sector, strategic metals such as Nickel and Palladium exhibit higher
average weights (Tesla–Nickel = 0.26; BYD–Nickel = 0.30; Tesla–Palladium = 0.27; BYD–
Palladium = 0.33), highlighting their substantial contribution to optimal hedging strategies.
This is consistent with their essential role in battery technologies and EV drivetrain systems.
In contrast, Gold and Copper register lower average allocations (e.g., Tesla–Gold = 0.08;
BYD–Gold = 0.10; Tesla–Copper = 0.11; BYD–Copper = 0.11), reaffirming their relatively
weak co-movement and stable, but limited, hedging influence. These patterns align with
previous findings in Table 6 and Fig. 8, which reveal high hedge ratios for Nickel and
Palladium and lower, more stable hedge ratios for Gold.

106

107

108

109

110

Fig. 9 Boxplots of Time-Varying Hedge Ratios Between Equity Stocks and Strategic/Precious Metals Across
Crisis and Normal Periods.
Note: This figure presents the distribution of dynamic hedge ratios derived from DCC-GARCH estimates between each
equity–metal pair across three market states: normal, COVID-19 (March–June 2020), and the Russia–Ukraine War (February
2022–December 2022). Elevated hedge ratios during crises indicate higher hedging costs and stronger co-movements due to
intensified market uncertainty.

For traditional automakers, optimal weights reveal that Nickel continues to play a
dominant hedging role, with average weights exceeding 0.50 in most cases (Toyota = 0.66;
Honda = 0.59; GM = 0.50; Ford = 0.50). These findings underscore the industrial dependence
of automakers on strategic metals for efficient vehicle design and emissions reduction.
Conversely, Gold consistently receives lower portfolio weights (Toyota = 0.33; Honda = 0.27;
GM = 0.21; Ford = 0.21), reinforcing its passive hedging function and low correlation with
automobile stock returns (Baur & Lucey, 2010).

111

Table 7 Summary statistics for the optimal weights across different equity indices and metals
Pair

Mean

Median

SD

Min

Max

Toyota/Copper

0.4287

0.4351

0.1355

0.0462

0.791

Toyota/Aluminium

0.4451

0.4428

0.1526

0.0399

0.9192

Toyota/Nickel

0.662

0.6883

0.1453

0.0835

0.9187

Toyota/Zinc

0.5539

0.5743

0.1442

0.0847

0.858

Toyota/Cobalt

0.2878

0.1761

0.3017

-0.0022

1.014

Toyota/Gold

0.3342

0.3379

0.0996

0.0756

0.6126

Toyota/Silver

0.6221

0.6332

0.131

0.1392

0.9207

Toyota/Platinum

0.5399

0.5404

0.1488

0.0777

0.8942

Toyota/Palladium

0.6783

0.6956

0.1321

0.1235

0.9491

Honda/Copper

0.3424

0.3377

0.117

0.0588

0.7172

Honda/Aluminium

0.3738

0.3607

0.146

0.0644

0.8425

Honda/Nickel

0.587

0.5934

0.1477

0.1704

0.8794

Honda/Zinc

0.4704

0.4726

0.1342

0.1553

0.8218

Honda/Cobalt

0.2561

0.1439

0.2807

0.0006

0.994

Honda/Gold

0.2719

0.2652

0.0652

0.109

0.4645

Honda/Silver

0.5533

0.5465

0.1059

0.2542

0.8435

Honda/Platinum

0.4646

0.4691

0.1239

0.1346

0.7472

Honda/Palladium

0.607

0.599

0.1328

0.2622

0.9707

GM/Copper

0.2542

0.2553

0.112

-0.0732

0.6769

GM/Aluminium

0.2875

0.2843

0.1142

-0.0112

0.7585

GM/Nickel

0.495

0.5202

0.146

-0.0374

0.7748

GM/Zinc

0.3802

0.3881

0.1346

-0.0244

0.6819

GM/Cobalt

0.2271

0.1174

0.2607

0.0004

0.986

GM/Gold

0.207

0.2023

0.0847

0.0257

0.4878

GM/Silver

0.46

0.4612

0.1292

0.0664

0.7566

GM/Platinum

0.3524

0.3613

0.1061

0.0481

0.5835

GM/Palladium

0.5154

0.524

0.1293

0.1896

0.8639

F/Copper

0.2702

0.2449

0.1348

-0.0179

0.8234

F/Aluminium

0.2948

0.2691

0.1375

0.0212

0.8059

F/Nickel

0.5018

0.518

0.1644

0.0023

0.8168

F/Zinc

0.387

0.3901

0.155

0.0179

0.7455

F/Cobalt

0.2361

0.125

0.2679

0.0004

0.9848

F/Gold

0.214

0.2045

0.0979

0.0247

0.472

F/Silver

0.4665

0.4726

0.145

0.1459

0.7865

F/Platinum

0.3613

0.3609

0.1063

0.1142

0.6462

F/Palladium

0.5242

0.5284

0.1277

0.1672

0.8142

Tesla/Copper

0.1117

0.0979

0.0663

0.0071

0.5452

Tesla/Aluminium

0.1307

0.1216

0.0677

-0.0007

0.4729

Tesla/Nickel

0.2602

0.2683

0.1038

0.0039

0.5697

Tesla/Zinc

0.1754

0.1662

0.082

-0.0025

0.4876

Tesla/Cobalt

0.1228

0.0385

0.1825

-0.003

0.9838

Tesla/Gold

0.0781

0.0702

0.0351

0.0101

0.209

112

Tesla/Silver

0.2314

0.2191

0.0924

0.0481

0.4933

Tesla/Platinum

0.1547

0.148

0.0735

0.0223

0.4884

Tesla/Palladium

0.2662

0.2517

0.1239

0.0359

0.725

BYD/Copper

0.1075

0.0911

0.0888

-0.0334

0.6291

BYD/Aluminium

0.1416

0.1284

0.0949

0.003

0.5953

BYD/Nickel

0.3005

0.2823

0.1594

-0.0099

0.6696

BYD/Zinc

0.2029

0.1878

0.1322

-0.0009

0.6467

BYD/Cobalt

0.1435

0.0365

0.218

-0.0954

1.1802

BYD/Gold

0.1037

0.0894

0.0553

0.0216

0.3067

BYD/Silver

0.285

0.2705

0.1162

0.0396

0.6494

BYD/Platinum

0.2202

0.201

0.1156

0.0084

0.5473

BYD/Palladium

0.3316

0.3253

0.1709

0.0191

0.8528

CELS/Copper

0.3264

0.3057

0.1742

-0.0437

0.8767

CELS/Aluminium

0.3513

0.3381

0.1675

-0.0106

0.8619

CELS/Nickel

0.5652

0.5856

0.1909

-0.0641

0.9118

CELS/Zinc

0.4497

0.4433

0.1931

0.0024

0.8718

CELS/Cobalt

0.2656

0.1446

0.2943

-0.0023

0.9997

CELS/Gold

0.2476

0.2394

0.1344

-0.024

0.572

CELS/Silver

0.5326

0.5541

0.1743

0.0507

0.9603

CELS/Platinum

0.4274

0.4171

0.169

0.0389

0.8174

CELS/Palladium

0.5942

0.6095

0.1659

0.0706

0.9499

SPGTCLEN/Copper

0.4909

0.4875

0.1597

0.0267

0.9488

SPGTCLEN/Aluminium

0.507

0.5257

0.1609

0.0237

0.9247

SPGTCLEN/Nickel

0.6993

0.7363

0.153

0.0456

0.9151

SPGTCLEN/Zinc

0.5987

0.6246

0.1582

0.0486

0.8872

SPGTCLEN/Cobalt

0.3379

0.244

0.3191

0.0006

1.0002

SPGTCLEN/Gold

0.3812

0.3803

0.1465

0.0354

0.9271

SPGTCLEN/Silver

0.6707

0.6972

0.142

0.1377

0.9886

SPGTCLEN/Platinum

0.5916

0.6041

0.1382

0.1621

0.9902

SPGTCLEN/Palladium

0.7093

0.7322

0.1375

0.1597

0.993

ECO/Copper

0.2843

0.2592

0.1658

-0.0691

0.8199

ECO/Aluminium

0.3084

0.2848

0.1612

-0.02

0.8799

ECO/Nickel

0.5228

0.5317

0.1943

-0.0636

0.8787

ECO/Zinc

0.4078

0.4064

0.1901

-0.0005

0.8228

ECO/Cobalt

0.2476

0.1257

0.2833

-0.0018

0.9979

ECO/Gold

0.2166

0.2122

0.123

-0.0154

0.5266

ECO/Silver

0.4867

0.4912

0.1747

0.0562

0.8992

ECO/Platinum

0.3777

0.3552

0.1574

0.0566

0.7716

ECO/Palladium

0.5467

0.5559

0.1664

0.0762

0.9354

Note: This table reports the summary statistics of the optimal portfolio weights between equity indices (EV stocks, traditional
automakers, and clean energy indices) and strategic/precious metals over the 2014–2024 period. The weights reflect the
proportion of equity in a two-asset minimum-variance portfolio, with the remainder invested in the corresponding metal.

113

In the clean energy sector, metals such as Nickel, Palladium, and Silver command the
highest average portfolio allocations across CELS, ECO, and SPGTCLEN. Notably,
SPGTCLEN–Palladium (0.71), SPGTCLEN–Nickel (0.70), and SPGTCLEN–Silver (0.67)
highlight the deep financial and technological linkage between metal inputs and renewable
energy firms. This is particularly relevant for energy storage, solar panel production, and wind
turbine components. Meanwhile, Gold continues to show moderate to low weights across clean
energy indices (e.g., ECO–Gold = 0.22; CELS–Gold = 0.25), consistent with its role as a lowcost, stable hedge rather than an active risk absorber.
Overall, the results of Table 7 support key thesis hypotheses. First, strategic metals tend
to exhibit higher optimal weights, especially during times of market stress, validating their
increasing financialization and integration into equity risk structures. Second, EV and clean
energy stocks demonstrate greater metal sensitivity than traditional automakers, reflected in
both higher hedge ratios and larger portfolio allocations to critical materials. Third, precious
metals—especially Gold—offer passive, stable hedging benefits, but are less integrated into
the volatility dynamics of these equity sectors. These findings emphasize the importance of
dynamic, sector-specific portfolio strategies for investors seeking to mitigate exposure to raw
material shocks, especially in periods of geopolitical or supply chain disruptions.
4.7 Robustness of Empirical Findings
The Fig. 10 displays the evolution of the Total Connectedness Index (TCI) among
strategic metals, precious metals, automaker stocks, and clean energy indices from Jan 2014 to
Jan 2024. The TCI is computed using different rolling window lengths (250, 300, 350, and 400
days) to assess the robustness of systemic connectedness estimates. The presence of sharp
spikes in early 2020 and early 2022 aligns with the onset of the COVID-19 pandemic and the
114

Russia–Ukraine conflict, respectively, indicating heightened volatility spillovers during
periods of global uncertainty. The consistency of TCI patterns across varying window sizes
reinforces the robustness of the empirical findings and confirms the reliability of the detected
interlinkages among strategic metals, precious metals, automakers, and clean energy indices.

Fig. 10 Robustness check based on total spillovers from Diebold and Yilmaz (2012). RW, rolling window.
Note: This figure presents the Total Connectedness Index (TCI) across strategic metals, precious metals, automaker stocks,
and clean energy indices over the period January 2014 to January 2024. The TCI is computed using multiple rolling window
lengths (250, 300, 350, and 400 days) to evaluate the robustness of spillover estimates. The stability of TCI patterns across
different windows confirms that the observed interdependencies are not sensitive to methodological choices, thereby
enhancing the credibility of the empirical result

Chapter Five: CONCLUSION
This thesis explores the dynamic relationships between strategic and precious metals
and equity markets, including electric vehicle (EV) manufacturers, traditional automakers, and
clean energy indices, using advanced econometric methods such as DCC-GARCH, VAR,
Granger causality, impulse response functions (IRF), and the Diebold and Yilmaz (2012)
volatility spillover index. The study covers the period from 2014 to 2024, including major
global shocks like the COVID-19 pandemic and the Russia–Ukraine conflict, to capture both
normal and crisis market dynamics.

115

The DCC-GARCH results confirm Hypothesis H1 by showing strong time-varying
correlations between strategic metals (e.g., copper, aluminum, nickel, cobalt, zinc) and equity
markets—particularly EV stocks and clean energy indices. These correlations became stronger
during crisis periods, highlighting the increased co-movement between commodity and equity
markets during uncertainty.
Granger causality and IRF analysis support Hypotheses H2 and H3. Strategic metals,
especially cobalt, nickel, and aluminum, significantly Granger-cause and trigger short-term
reactions in EV stock returns, while their influence on traditional automakers is weaker and
more temporary. This confirms that strategic metals are more closely linked to EV markets
than to conventional car manufacturers.
The Diebold-Yilmaz connectedness analysis, with a Total Connectedness Index (TCI)
of 49.4%, shows moderate to high volatility spillovers across markets. Strategic metals were
more connected to EV and clean energy sectors, supporting Hypothesis H4, while precious
metals had broader but less sector-specific impact.
Hedge ratio and hedge effectiveness results confirm Hypothesis H5. Strategic metals
like copper and aluminum offered more responsive and effective hedging for EV and clean
energy stocks, especially during crisis periods. In contrast, gold and silver provided more stable
but passive hedging benefits, with lower sensitivity to equity market shocks. Platinum, due to
its high volatility, proved to be a less effective hedge overall.
Optimal portfolio weights further support these findings. During periods of high
uncertainty, investors should increase their allocation to strategic metals when managing EV
and clean energy exposures, while gold remains useful as a general hedge during systemic risk.

116

Overall, this thesis makes three key contributions:
1. It bridges clean energy finance with commodity markets by modeling both strategic
and precious metals in a dynamic portfolio and volatility framework.
2. It provides robust evidence, across different models and time horizons—on how metals
contribute to financial risk and hedging performance.
3. It offers insights for investors and policymakers on the growing importance of strategic
metals in green finance, supply chain risk, and energy transition planning.
In summary, the study shows that strategic and precious metals play different but
complementary roles in financial markets. Strategic metals are increasingly integrated into the
risk structure of EV and clean energy stocks, while precious metals continue to serve as safehaven assets. As markets evolve due to technological advances and geopolitical shocks,
understanding these relationships is essential for building resilient investment and policy
strategies.
BIBLIOGRAPHY
Alekseev, O., Janda, K., Petit, M., & Zilberman, D. (2024). Return and volatility spillovers
between the raw material and electric vehicles markets. Energy Economics, 137,
107808. https://doi.org/10.1016/j.eneco.2024.107808
Amândio, H., Dias, R., Galvão, R., Alexandre, P., Gonçalves, S., & Leote, F. (2025). Are
Precious Metals Hedging Assets for Clean Energy Indices? Journal of Ecohumanism,
4(1), Article 1. https://doi.org/10.62754/joe.v4i1.6388

117

Arouri, M. E. H., Lahiani, A., & Nguyen, D. K. (2011). Return and volatility transmission
between world oil prices and stock markets of the GCC countries. Economic
Modelling, 28(4), 1815–1825. https://doi.org/10.1016/j.econmod.2011.03.012
Baars, J., Domenech, T., Bleischwitz, R., Melin, H. E., & Heidrich, O. (2021). Circular
economy strategies for electric vehicle batteries reduce reliance on raw materials.
Nature Sustainability, 4(1), 71–79. https://doi.org/10.1038/s41893-020-00607-0
Bahini, Y., Mushtaq, R., & Bahoo, S. (2024). Global energy transition: The vital role of
cobalt in renewable energy. Journal of Cleaner Production, 470, 143306.
https://doi.org/10.1016/j.jclepro.2024.143306
Banerjee, A. K., & Pradhan, H. (2024). Did Precious Metals Serve as Hedge and Safe-haven
Alternatives to Equity During the COVID-19 Pandemic: New Insights Using a
Copula-based Approach. Journal of Emerging Market Finance, 23(4), 399–423.
https://doi.org/10.1177/09726527241251515
Baruník, J., & Křehlík, T. (2018). Measuring the Frequency Dynamics of Financial
Connectedness and Systemic Risk*. Journal of Financial Econometrics, 16(2), 271–
296. https://doi.org/10.1093/jjfinec/nby001
Bastianin, A., Li, X., & Shamsudin, L. (2025). Forecasting the Volatility of Energy Transition
Metals (arXiv:2501.16069). arXiv. https://doi.org/10.48550/arXiv.2501.16069
Baur, D. G., & Lucey, B. M. (2010a). Is Gold a Hedge or a Safe Haven? An Analysis of
Stocks, Bonds and Gold. Financial Review, 45(2), 217–229.
https://doi.org/10.1111/j.1540-6288.2010.00244.x

118

Baur, D. G., & Lucey, B. M. (2010b). Is Gold a Hedge or a Safe Haven? An Analysis of
Stocks, Bonds and Gold. Financial Review, 45(2), 217–229.
https://doi.org/10.1111/j.1540-6288.2010.00244.x
Baur, D. G., & McDermott, T. K. (2010). Is gold a safe haven? International evidence.
Journal of Banking & Finance, 34(8), 1886–1898.
https://doi.org/10.1016/j.jbankfin.2009.12.008
Baur, D. G., & Smales, L. A. (2018). Gold and Geopolitical Risk (SSRN Scholarly Paper
3109136). Social Science Research Network. https://doi.org/10.2139/ssrn.3109136
Beckmann, J., Berger, T., & Czudaj, R. (2019). Gold price dynamics and the role of
uncertainty. Quantitative Finance, 19(4), 663–681.
https://doi.org/10.1080/14697688.2018.1508879
Bekaert, G., & Harvey, C. R. (1995). Time-Varying World Market Integration. The Journal of
Finance, 50(2), 403–444. https://doi.org/10.1111/j.1540-6261.1995.tb04790.x
Billio, M., Getmansky, M., Lo, A. W., & Pelizzon, L. (2012). Econometric measures of
connectedness and systemic risk in the finance and insurance sectors. Journal of
Financial Economics, 104(3), 535–559. https://doi.org/10.1016/j.jfineco.2011.12.010
Brooks, C. (2019, March 27). Introductory Econometrics for Finance. Higher Education
from Cambridge University Press; Cambridge University Press.
https://doi.org/10.1017/9781108524872
Cagli, E. C. (2023). The volatility spillover between battery metals and future mobility
stocks: Evidence from the time-varying frequency connectedness approach.
Resources Policy, 86, 104144. https://doi.org/10.1016/j.resourpol.2023.104144

119

Campbell, J. Y., & Shiller, R. J. (1988). Stock Prices, Earnings, and Expected Dividends. The
Journal of Finance, 43(3), 661–676. https://doi.org/10.1111/j.15406261.1988.tb04598.x
Caporale, G. M., & Gil-Alana, L. A. (2023). Gold and silver as safe havens: A fractional
integration and cointegration analysis. PLOS ONE, 18(3), e0282631.
https://doi.org/10.1371/journal.pone.0282631
Chang, Y., Fang, Z., & Hamori, S. (2017). Volatility and Causality in Strategic Commodities:
Characteristics, Myth and Evidence. International Journal of Economics and
Finance, 9(8), Article 8. https://doi.org/10.5539/ijef.v9n8p162
Chen, N.-F., Roll, R., & Ross, S. A. (1986). Economic Forces and the Stock Market. The
Journal of Business, 59(3), 383–403. https://www.jstor.org/stable/2352710
Chen, Y., Zhu, X., & Chen, J. (2022). Spillovers and hedging effectiveness of non-ferrous
metals and sub-sectoral clean energy stocks in time and frequency domain. Energy
Economics, 111, 106070. https://doi.org/10.1016/j.eneco.2022.106070
Clancey-Shang, D., & Fu, C. (2023). The Russia–Ukraine conflict and foreign stocks on the
US market. The Journal of Risk Finance, 24(1), 6–23. https://doi.org/10.1108/JRF07-2022-0179
Clancey-Shang, D., & Fu, C. (2024). CSR disclosure, political risk and market quality:
Evidence from the Russia-Ukraine conflict. Global Finance Journal, 60, 100938.
https://doi.org/10.1016/j.gfj.2024.100938
Cont, R. (2001). Empirical properties of asset returns: Stylized facts and statistical issues.
Quantitative Finance, 1(2), 223–236. https://doi.org/10.1080/713665670

120

Coudert, V. & Raymond, H. (2011). Gold and financial assets: Are there any safe havens in
bear markets? ResearchGate, 31(2), 1613–1622.
https://www.researchgate.net/publication/227352893_Gold_and_financial_assets_Are
_there_any_safe_havens_in_bear_markets
Diebold, F. X., & Yilmaz, K. (2012). Better to give than to receive: Predictive directional
measurement of volatility spillovers. International Journal of Forecasting, 28(1), 57–
66.
Ding, S., Cui, T., Zheng, D., & Du, M. (2021). The effects of commodity financialization on
commodity market volatility. Resources Policy, 73, 102220.
https://doi.org/10.1016/j.resourpol.2021.102220
Dinh, T., Goutte, S., Nguyen, D. K., & Walther, T. (2022). Economic drivers of volatility and
correlation in precious metal markets. Journal of Commodity Markets, 28, 100242.
https://doi.org/10.1016/j.jcomm.2021.100242
Dutta, A. (2019). Impact of silver price uncertainty on solar energy firms. Journal of Cleaner
Production, 225, 1044–1051.
Dutta, A., Bouri, E., Das, D., & Roubaud, D. (2020). Assessment and optimization of clean
energy equity risks and commodity price volatility indexes: Implications for
sustainability. Journal of Cleaner Production, 243, 118669.
Elie, B., Naji, J., Dutta, A., & Uddin, G. S. (2019). Gold and crude oil as safe-haven assets
for clean energy stock indices: Blended copulas approach. Energy, 178, 544–553.
Engle, R. (2002a). Dynamic Conditional Correlation. Journal of Business & Economic
Statistics. https://doi.org/10.1198/073500102288618487

121

Engle, R. (2002b). Dynamic Conditional Correlation: A Simple Class of Multivariate
Generalized Autoregressive Conditional Heteroskedasticity Models. Journal of
Business & Economic Statistics, 20(3), 339–350.
https://www.jstor.org/stable/1392121
Erb, C. B., & Harvey, C. R. (2006). The Tactical and Strategic Value of Commodity Futures
(SSRN Scholarly Paper 650923). Social Science Research Network.
https://doi.org/10.2139/ssrn.650923
Erdoğan, S., Gedikli, A., Çevik, E. İ., Erdoğan, F., & Çevik, E. (2022). Precious metals as
safe-haven for clean energy stock investment: Evidence from nonparametric Granger
causality in distribution test. Resources Policy, 79, 102945.
Fernanda Soares, A., Giovinazzo Spers, R., & de Oliveira Santos Jhunior, R. (2025).
Projection of global copper demand in the context of energy transition. Resources
Policy, 103, 105567. https://doi.org/10.1016/j.resourpol.2025.105567
Fu, C., Huang, Q., Lin, M., & Tahsin, S. (2025). Rising Tides, Sinking Approval Rates:
Examining SLR Risk and Mortgage Credit Access. European Financial Management,
n/a(n/a). https://doi.org/10.1111/eufm.12543
Fu, C., Jacoby, G., Lin, N., & Lu, L. (2025). The Value of Investor Sophistication. Financial
Management, n/a(n/a). https://doi.org/10.1111/fima.12497
Ghosh, B., Pham, L., Teplova, T., & Umar, Z. (2023). COVID-19 and the quantile
connectedness between energy and metal markets. Energy Economics, 117, 106420.
https://doi.org/10.1016/j.eneco.2022.106420
Global EV Outlook 2023 – Analysis. (2023, April 26). IEA.
https://www.iea.org/reports/global-ev-outlook-2023

122

Gong, X., Fu, C., Huang, Q., & Lin, M. (2022). International political uncertainty and
climate risk in the stock market. Journal of International Financial Markets,
Institutions and Money, 81, 101683. https://doi.org/10.1016/j.intfin.2022.101683
Gong, X., Guan, K., Chen, L., Liu, T., & Fu, C. (2021). What drives oil prices? — A Markov
switching VAR approach. Resources Policy, 74, 102316.
https://doi.org/10.1016/j.resourpol.2021.102316
Gong, X., & Jia, G.-Z. (2024). Exploring the Cross-Correlations between Tesla Stock Price,
New Energy Vehicles and Oil Prices: A Multifractal and Causality Analysis.
Fluctuation and Noise Letters, 23(03), 2450024.
https://doi.org/10.1142/S021947752450024X
Gong, X., Song, Y., Fu, C., & Li, H. (2023). Climate risk and stock performance of fossil fuel
companies: An international analysis. Journal of International Financial Markets,
Institutions and Money, 89, 101884. https://doi.org/10.1016/j.intfin.2023.101884
Graedel, T. E., Harper, E. M., Nassar, N. T., Nuss, P., & Reck, B. K. (2015). Criticality of
metals and metalloids. Proceedings of the National Academy of Sciences, 112(14),
4257–4262. https://doi.org/10.1073/pnas.1500415112
Granger, C. W. J. (1969). Investigating Causal Relations by Econometric Models and Crossspectral Methods. Econometrica, 37(3), 424–438. https://doi.org/10.2307/1912791
Gustafsson, R., Dutta, A., & Bouri, E. (2022). Are energy metals hedges or safe havens for
clean energy stock returns? Energy, 244, 122708.
Hamilton, J. D. (1994). Time series analysis (1st edition). Princeton university press.
Hasbrouck, J. (1991). Measuring the Information Content of Stock Trades. The Journal of
Finance, 46(1), 179–207. https://doi.org/10.1111/j.1540-6261.1991.tb03749.x

123

Hobson, P. (2022, March 7). Palladium propelled to record highs by Russia supply concerns.
Reuters. https://www.reuters.com/business/palladium-propelled-record-highs-byrussia-supply-concerns-2022-03-07/
IEA. (2021, May 5). The Role of Critical Minerals in Clean Energy Transitions – Analysis.
https://www.iea.org/reports/the-role-of-critical-minerals-in-clean-energy-transitions
Jégourel, Y. (2021). Financialization of the Minerals and Metals Market: Origin, Challenges
and Prospects. In Mineral Resources Economics 1 (pp. 35–52). John Wiley & Sons,
Ltd. https://doi.org/10.1002/9781119850861.ch2
Kang, W., Tang, K., & Wang, N. (2023). Financialization of commodity markets ten years
later. Journal of Commodity Markets, 30, 100313.
https://doi.org/10.1016/j.jcomm.2023.100313
Kojić, M., Mitić, P., & Minović, J. (2023). Gold and Sustainable Stocks in the US and EU:
Nonlinear Analysis Based on Multifractal Detrended Cross-Correlation Analysis and
Granger Causality. Fractal and Fractional, 7(10), Article 10.
https://doi.org/10.3390/fractalfract7100738
Koop, G., Pesaran, M. H., & Potter, S. M. (1996). Impulse response analysis in nonlinear
multivariate models. Journal of Econometrics, 74(1), 119–147.
Kroner, K. F., & Ng, V. K. (1998). Modeling asymmetric comovements of asset returns. The
Review of Financial Studies, 11(4), 817–844.
Lahiani, A., Mefteh-Wali, S., & Vasbieva, D. G. (2021). The safe-haven property of precious
metal commodities in the COVID-19 era. Resources Policy, 74, 102340.
https://doi.org/10.1016/j.resourpol.2021.102340

124

Lander, L., Cleaver, T., Rajaeifar, M. A., Nguyen-Tien, V., Elliott, R. J. R., Heidrich, O.,
Kendrick, E., Edge, J. S., & Offer, G. (2021). Financial viability of electric vehicle
lithium-ion battery recycling. iScience, 24(7), 102787.
https://doi.org/10.1016/j.isci.2021.102787
Li, H., Clancey-Shang, D., Fu, C., & Li, T. (2025). Does the world need more traditional
energy? A comparative analysis of ESG activities, free cash flow, and capital market
implications. International Review of Financial Analysis, 99, 103919.
https://doi.org/10.1016/j.irfa.2025.103919
Liu, X., Li, S., Fu, C., Gong, X., & Fan, C. (2024). The oil price plummeted in 2014–2015: Is
there an effect on Chinese firms’ labour investment? International Journal of Finance
& Economics, 29(1), 943–960. https://doi.org/10.1002/ijfe.2715
Lucey, B. M., & Li, S. (2015). What precious metals act as safe havens, and when? Some US
evidence. Applied Economics Letters.
https://www.tandfonline.com/doi/abs/10.1080/13504851.2014.920471
Lütkepohl, H. (2005). New Introduction to Multiple Time Series Analysis. Springer.
https://doi.org/10.1007/978-3-540-27752-1
Mayer, H., Rathgeber, A., & Wanner, M. (2017). Financialization of metal markets: Does
futures trading influence spot prices and volatility? Resources Policy, 53, 300–316.
https://doi.org/10.1016/j.resourpol.2017.06.011
Mensi, W., Al Rababa’a, A. R., Alomari, M., Vo, X. V., & Kang, S. H. (2022). Dynamic
frequency volatility spillovers and connectedness between strategic commodity and
stock markets: US-based sectoral analysis. Resources Policy, 79, 102976.
https://doi.org/10.1016/j.resourpol.2022.102976

125

Mensi, W., Rehman, M. U., & Vo, X. V. (2020). Spillovers and co-movements between
precious metals and energy markets: Implications on portfolio management.
Resources Policy, 69, 101836. https://doi.org/10.1016/j.resourpol.2020.101836
Nguyen, R. T., Eggert, R. G., Severson, M. H., & Anderson, C. G. (2021). Global
Electrification of Vehicles and Intertwined Material Supply Chains of Cobalt, Copper
and Nickel. Resources, Conservation and Recycling, 167, 105198.
https://doi.org/10.1016/j.resconrec.2020.105198
Pakrooh, P., & Manera, M. (2024). Causality, Connectedness, and Volatility pass-through
among Energy-Metal-Stock-Carbon Markets: New Evidence from the EU. Resources
Policy, 99, 105408. https://doi.org/10.1016/j.resourpol.2024.105408
Pesaran, H. H., & Shin, Y. (1998). Generalized impulse response analysis in linear
multivariate models. Economics Letters, 58(1), 17–29.
Pham, L., & Hsu, K.-C. (2025). Metals of the future in a world in crisis: Geopolitical
disruptions and the cleantech metal industry. Energy Economics, 141, 108004.
https://doi.org/10.1016/j.eneco.2024.108004
Rajwani, S., Tiwari, A., Yadav, M., & Sharma, S. (2023). Dynamic Linkages of Energy
Commodities with Bullion and Metal Market: Evidence of Portfolio Hedging.
American Business Review, 26(1). https://doi.org/10.37625/abr.26.1.148-179
Reboredo, J. C. (2013). Is gold a safe haven or a hedge for the US dollar? Implications for
risk management. Journal of Banking & Finance, 37(8), 2665–2676.
https://doi.org/10.1016/j.jbankfin.2013.03.020

126

S, J. S., Kayal, P., & Balasubramanian, G. (2022). Can Equity be Safe-haven for Investment?
Journal of Emerging Market Finance, 21(1), 32–63.
https://doi.org/10.1177/09726527211068411
Sadorsky, P. (2012). Correlations and volatility spillovers between oil prices and the stock
prices of clean energy and technology companies. Energy Economics, 34(1), 248–
255.
Sari, R., Hammoudeh, S., & Soytas, U. (2010). Dynamics of oil price, precious metal prices,
and exchange rate. Energy Economics, 32(2), 351–362.
https://doi.org/10.1016/j.eneco.2009.08.010
Schwert, G. W. (1989). Why Does Stock Market Volatility Change Over Time? The Journal
of Finance, 44(5), 1115–1153. https://doi.org/10.1111/j.1540-6261.1989.tb02647.x
Sims, C. A. (1980). Macroeconomics and Reality. Econometrica, 48(1), 1–48.
https://doi.org/10.2307/1912017
Sims, C.-I. A. (1980). ECONO MI ETRICA. Econometrica, 48(1), 1–48.
Sovacool, B. K., Ali, S. H., Bazilian, M., Radley, B., Nemery, B., Okatz, J., & Mulvaney, D.
(2020). Sustainable minerals and metals for a low-carbon future. Science, 367(6473),
30–33. https://doi.org/10.1126/science.aaz6003
Swarup, S., & Kushwaha, G. S. (2023). Nickel and Cobalt Price Volatility Forecasting Using
a Self-Attention-Based Transformer Model. Applied Sciences, 13(8), Article 8.
https://doi.org/10.3390/app13085072
Tan, X., Wang, X., Ma, S., Wang, Z., Zhao, Y., & Xiang, L. (2022). COVID-19 Shock and
the Time-Varying Volatility Spillovers Among the Energy and Precious Metals

127

Markets: Evidence From A DCC-GARCH-CONNECTEDNESS Approach. Frontiers
in Public Health, 10. https://doi.org/10.3389/fpubh.2022.906969
Tang, K., & and Xiong, W. (2012). Index Investment and the Financialization of
Commodities. Financial Analysts Journal, 68(6), 54–74.
https://doi.org/10.2469/faj.v68.n6.5
The Role of Critical Minerals in Clean Energy Transitions – Analysis. (2021, May 5). IEA.
https://www.iea.org/reports/the-role-of-critical-minerals-in-clean-energy-transitions
van den Brink, S., Kleijn, R., Sprecher, B., & Tukker, A. (2020). Identifying supply risks by
mapping the cobalt supply chain. Resources, Conservation and Recycling, 156,
104743. https://doi.org/10.1016/j.resconrec.2020.104743
Westra, R. (2024). The Rare Earths Era: Strategic Metals Dependency & World Order: Juan
Manuel Chomón. Atlanta: Clarity Press, 2024. Journal of Contemporary Asia, 0(0),
1–4. https://doi.org/10.1080/00472336.2025.2486589
Włodarczyk, B., & Szturo, M. (2018). Financialization of Commodity Markets. In K. Jajuga,
H. Locarek-Junge, & L. T. Orlowski (Eds.), Contemporary Trends and Challenges in
Finance (pp. 99–108). Springer International Publishing. https://doi.org/10.1007/9783-319-76228-9_10
World Bank. (2022). Precious Metal Prices Pressured by Rising Interest Rates and Weaker
Economic Activity. https://blogs.worldbank.org/en/opendata/precious-metal-pricespressured-rising-interest-rates-and-weaker-economic-activity
World Health Organization. (2020). Timeline: WHO’s COVID-19 response.
https://www.who.int/emergencies/diseases/novel-coronavirus-2019/interactivetimeline

128

Xing, X., Xu, Z., Chen, Y., Ouyang, W., Deng, J., & Pan, H. (2023). The impact of the
Russia–Ukraine conflict on the energy subsector stocks in China: A network-based
approach. Finance Research Letters, 53, 103645.
https://doi.org/10.1016/j.frl.2023.103645
Yahya, M., Ghosh, S., Kanjilal, K., Dutta, A., & Uddin, G. S. (2020). Evaluation of crossquantile dependence and causality between non-ferrous metals and clean energy
indexes. Energy, 202, 117777. https://doi.org/10.1016/j.energy.2020.117777
Yin, L., & Cao, H. (2024). Financialization of commodity markets: New evidence from
temporal and spatial domains. Journal of Futures Markets, 44(8), 1357–1382.
https://doi.org/10.1002/fut.22514

129