Safe Haven Ability of Energy and Agricultural

Full text

Turn on search term navigation

This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1. Introduction

When investing in financial markets, investors often raise concerns about numerous market-related uncertainties and risks [1–6]. As a result, they look for suitable investments that offer hedging and safe haven possibilities against threats from financial markets, disruptive economic policy shifts, and financial crises such as the COVID-19 pandemic, the Russia–Ukraine military conflict, and the Silicon Valley Bank (SVB) collapse.

For instance, the COVID-19 crisis and its rapid spread globally have forced many countries to put in action rigid measures (such as isolations, shutdowns, and travel prohibitions), significantly limiting economic activities [2, 3, 7, 8]. Certainly, ever since the emergence of the COVID-19 outbreak, financial markets have encountered major instability and unanticipated declines in returns [3, 8, 9], particularly those of the G7 countries [10, 11]. Rehman et al. [12] found a significant correlation between the number of cases and deaths related to the COVID-19 pandemic and the G7 stock markets. Awan et al. [13] state that the COVID-19 pandemic increases volatility in all the G7 stock markets. Mobin et al. [14] affirm that negative news regarding the pandemic induced greater volatility when compared to positive news across all G7 stock markets. Belhouchet and Amara [15] investigate the response of the G7 stock markets to the side effects of the COVID-19 pandemic, revealing its adverse impact on the G7 stock market performance and the trading environments of these economies.

Banking stocks, on the other hand, are not immune from this health crisis. For instance, Matos et al. [16] study how financial stress spread among G7 banking indices before and during the COVID-19 pandemic. They found that the pandemic worsened the risk propagation within these banking systems. Notably, there was significant contagion between the banking systems of Italy and France, which were heavily affected by COVID-19 fatalities, whereas minimal contagion was observed between Germany and Japan, countries less impacted by COVID-19 fatalities. The same results are also detected by Matos, da Silva, and Costa [17]. Jeris and Nath [18] examine the influence of the COVID-19 pandemic on the performance of the bank indices in the United States. They discovered that the US bank indices were greatly affected by the rising number of COVID-19 cases in the country. Mirzaei, Saad, and Emrouznejad [19] investigate the performance of Islamic and traditional banking stocks during the COVID-19 pandemic. Their results indicate that Islamic banks showed approximately 10%–13% greater stock returns compared to conventional banks.

After the profound consequences of the COVID-19 pandemic, the world encountered a fresh obstacle with the start of the Russian–Ukrainian military confrontation on the 24^th of February, 2022, and it was Europe’s greatest military assault since World War II. The assault had a significant influence on financial markets [8, 20–22]. Most stock indexes have declined in value during this crisis, and many commodities’ prices have risen sharply [23]. Also, banks, enterprises, monetary transactions, exports, and imports were all influenced by this conflict due to the different economic sanctions imposed on Russia in an attempt to cripple its economy and make it stop the war [23]. Alam et al. [24] assert that the conflict between Russia and Ukraine had a major effect on the transmission of volatility between commodities and the G7 stock markets. Ahmed, Hasan, and Kamal [25] found that European stock markets reacted negatively to the Russia–Ukraine crisis due to increased political uncertainty, geographical proximity, and the impact of recent sanctions imposed on Russia. Boubaker et al. [26] discovered that worldwide stock market indices suffered negative returns following the Russian invasion of Ukraine in 2022. When considering the banking system, Martins, Correia, and Gouveia [27] examine the impact of the military conflict between Russia and Ukraine on the top 100 European banks. They found that these banks experienced a negative reaction in their stock prices at the outset and throughout the conflict. Moreover, their research reveals a heightened negative reaction in the stock market for banks based in Russia and for international banks engaged with the Russian market. Boubaker et al. [22] examine how the banking sector reacted to the Russia–Ukraine conflict. They found that on the day of the attack, there was a decrease of about 1% in the returns of worldwide bank assets, indicating a more noticeable effect on bank stocks in contrast to the broader stock market. Particularly, bank stocks in Europe, Asia, and North America saw the most significant decline.

More recently, on the 10^th of March 2023, the world observed the most considerable financial crisis since the 2008 global financial crisis, the collapse of one of the most prominent financial institutions in the United States, the SVB. Such unexpected downfall of the SVB considerably impacted global financial markets, creating disruptions in the financial system [28–31] and affecting international banks [32]. Pandey et al. [28] asserted that the downfall of the SVB negatively impacted global stock markets, notably in developed economies more than in emerging ones. Aharon, Ali, and Naved [30] stated that worldwide stock markets exhibited negative reactions to the SVB turmoil. Manda [33] examines the downfall of the SVB in the United States and its consequences for the international banking system, indicating the presence of risk transmission because of the global interconnectedness of financial services. Corbet and Larkin [34] examine the SVB collapse and its effects on the worldwide financial system. Their findings suggest that the bank’s collapse not only prompted worries regarding the transmission of contagion to other banks but also inflicted damage on global markets, highlighting the interconnectedness of the international banking system.

The ongoing health and political turmoil as well as the recent American banking crisis demonstrate a heightened degree of complexity compared to previous financial crises. Certainly, following these crises, numerous financial markets have encountered significant instability and unexpected declines in returns [8, 9, 15–17, 20–22, 30, 34]. During such periods, investors employ tactics to mitigate portfolio losses, including seeking refuge in diversifying, hedging, and safe haven assets.

In fact, Baur and Lucey [35] were the first to define hedging, diversifying, and safe haven assets. They described hedging assets as financial instruments that help investors safeguard their portfolios under normal market conditions. Diversifying assets, on the other hand, are financial instruments that allow investors to lower portfolio risk and boost diversification in both stable and turbulent times. Meanwhile, safe haven assets are financial instruments that exhibit either no correlation or negative correlation with another asset or portfolio during periods of severe stock market downturns. Unlike a hedge, investors seek a safe haven asset only during periods of market distress. The asset maintains its value during such times and therefore functions as a refuge [35–37]. In this manner, safe haven assets can aid investors in forming a portfolio that reduces adverse market risk during periods of turmoil, such as the COVID-19 outbreak, the Russia–Ukraine war, and the SVB collapse.

Most of the previous research examining the safe haven ability of financial assets focused mostly on the safe haven ability of gold or other precious metals [2, 3, 8, 11, 35, 36, 38–40]. Other works focused on the safe haven ability of digital assets, namely, cryptocurrencies [2, 3, 11, 41–44]. In our paper, we aim to explore several potential safe haven assets that are typically overlooked in existing literature, such as energy (crude oil and natural gas) and agricultural (wheat) commodities, and subsequently, we empirically assess their efficacy in mitigating risks associated with the G7 stock market indices and banking sector stock indices during periods of crises.

The choice of crude oil, natural gas, and wheat as commodities for diversifying, hedging, and safe haven purposes in this study is justified by their significant economic impact [45–47], historical and contemporary relevance [48, 49], and the availability of reliable high-frequency data [50]. Crude oil and natural gas, as pivotal energy commodities, wield substantial influence on global markets, while wheat is a critical agricultural commodity essential for food security. These commodities have demonstrated notable price movements in response to geopolitical events, economic crises, and supply–demand shocks, making them appropriate for examining hedging and safe haven properties [3, 51–53].

Furthermore, recent events such as the COVID-19 pandemic, Russia–Ukraine war, and SVB collapse have directly impacted these commodities, underscoring their importance in understanding market dynamics during crises [54]. Oil prices, for instance, fluctuate differently from other financial markets, providing useful hedging and diversification benefits [55]. For many years, natural gas prices were closely related to those of crude oil. However, natural gas prices have recently become more independent [56]. These prices are essentially determined by supply and demand dynamics, which are impacted by factors such as weather and inventory levels. The changing pricing trends in natural gas emphasize its distinct market behavior and potential for hedging and safe haven tactics [56].

In addition, the commodity market, particularly for agricultural commodities such as wheat, has seen increased trader interest, functioning similarly to financial assets for investment and serving as an appealing diversification tool and risk hedge due to their low or negative association with stock market assets [57, 58]. Wheat, in particular, holds a paramount position in global agriculture due to its role as a staple food for a large portion of the world’s population [59]. Its importance is further underscored by Russia’s emergence as the world’s largest wheat exporter, making Russian grain exports vital to regional and global food security. The significance of wheat extends beyond its nutritional value. It is also a key commodity in global trade, subject to price volatility driven by different events [59]. The financialization of commodities has further accelerated investment in commodity derivatives, with portfolio managers favoring them for their low or negative correlations with traditional assets, the tendency to move favorably with inflation, and the potential for risk-adjusted returns comparable to stocks [60].

Based on the wavelet coherence approach, our results reveal distinctive correlations between commodities (WTI, natural gas, and wheat) and G7 stock market indices during the three crises of the COVID-19 outbreak, the Russia–Ukraine war, and the SVB collapse. WTI initially acts as a diversifier, transitioning into a safe haven during the SVB collapse. Gas displays positive and negative correlations prepandemic, emerging as a strong safe haven during COVID-19, particularly for certain European indices. Wheat starts as a weak diversifier, evolving into a robust safe haven during crises. Regarding G7 banking sector stock indices, WTI shows diversification potential and safe haven status, natural gas primarily serves as a safe haven, and wheat functions as both a diversifier and a strong safe haven under different market conditions.

Our paper significantly contributes to the existing literature by expanding the investigative range beyond conventional safe haven assets such as gold and cryptocurrencies and examines the safe haven capabilities of frequently neglected commodities, particularly energy (crude oil and natural gas) and agricultural (wheat) commodities. By employing the wavelet coherence approach, our study empirically assesses their effectiveness in mitigating risks associated not only with the G7 stock market indices but also with the G7 banking sector stock indices.

Indeed, we concentrate in our study on the G7 stock market indices for the reason that they are recognized as the world’s most substantial and influential economies [8]. Consequently, fluctuations in their stock markets echo internationally, influencing investor sentiment and market dynamics worldwide. Moreover, the G7 nations are commonly regarded as key players in global financial markets, with their stock indices serving as vital indicators of economic health and market stability [8]. On the other hand, the focus on the banking sector index is due to its critical role in the financial system and its heightened sensitivity during periods of economic stress, especially within the G7 countries. Specifically, the recent collapse of SVB highlighted the vulnerability and systemic importance of the banking sector [28, 32, 33]. Through the examination of commodities’ safe haven capabilities concerning these prominent stock markets, our investigation provides valuable perspectives into the broader implications of commodity investment strategies.

Our empirical findings provide new insights into the safe haven potential of energy and agricultural commodities during times of crisis, going beyond the boundaries of traditional safe haven assets (gold and cryptocurrencies). Moreover, we give a detailed examination of commodity performance under varied market settings by examining three major crises: the COVID-19 pandemic, the Russian–Ukrainian armed conflict, and the SVB collapse. This improves our understanding of portfolio diversification and risk management measures, providing investors with meaningful information for navigating volatile financial environments successfully. Our study also offers insights into the diversification benefits and protective capabilities of these essential commodities against the volatility of G7 stock market indices and banking sector stock indices during crises.

In essence, our paper fills a gap in the existing literature by providing empirical evidence of the dynamic correlation between commodities and financial markets during times of turmoil, as well as shedding light on the often-overlooked safe haven characteristics of energy and agricultural commodities, thus expanding the empirical foundation for modern portfolio management policies.

The remainder of our paper is structured as follows. Section 2 presents the literature review. Section 3 presents the data and descriptive statistics. Section 4 explores the empirical methodology. Section 5 presents the empirical results and discussion. Finally, Section 6 concludes the study.

2. Literature Review

Previous studies have thoroughly examined the safe haven attributes of various assets, concentrating primarily on gold [3, 35, 36], bonds [61–63], currencies [64–66], and cryptocurrencies [11, 40, 44, 67, 68]. While gold has consistently demonstrated qualities of a safe haven, the effectiveness of other commodities in offering comparable risk-mitigating advantages is still a topic of debate [5]. Moreover, research has delved into the correlation between commodities and equity markets, analyzing their behavior across various market conditions. Nonetheless, limited research has focused on the safe haven potential of energy and agricultural commodities concerning stock markets.

By employing the DCC, ADCC, and GO-GARCH techniques, Basher and Sadorsky [69] conclude that oil emerges as the most effective hedging tool for emerging stock markets when contrasted with other commodities. Bouoiyour, Selmi, and Wohar [70] investigate the safe haven function of crude oil for the US stock indices amid phases of political uncertainty. Their findings suggest that oil can serve as a reliable refuge during political risks. Jeribi and Snene-Manzli [3] assess the hedging and safe haven ability of crude oil for Tunisian stock prices during the COVID-19 pandemic. Their analysis illustrates its function as a safe haven asset. Kyriazis [54] investigates the impact of the Russian–Ukrainian military conflict on the performance of energy and agricultural commodities. Their results support commodities’ safe haven ability, suggesting that natural gas and wheat are the most profitable options.

Bandhu Majumder [51] evaluates the hedging and safe haven characteristics of gold, cryptocurrencies, and commodities for Indian equities amid the COVID-19 pandemic. The findings indicate that only crude oil, natural gas, and metals possess safe haven attributes for these equities. Ghorbel, Frikha, and Manzli [11] propose that WTI serves as a diversifying asset for the G7 stock market indices during the COVID-19 pandemic. Using a cross-quantilogram approach, Ji, Zhang, and Zhao [71] discovered that agricultural commodities serve as robust safe havens for global financial markets during the COVID-19 health crisis. Diaconaşu, Mehdian, and Stoica [72] investigate the repercussions of the Russia–Ukraine military conflict on worldwide commodities and stock markets. Their findings reveal that oil stands out as the sole asset offering safe haven features for investors at the onset of the conflict. Wang et al. [73] examine the hedging and safe haven ability of bitcoin, gold, and commodities against uncertainties in stock markets. Their results revealed that, even though all assets offer weak hedging abilities, commodities outperform bitcoin and gold as safe havens against stock markets during periods of turmoil. Hasan et al. [52] examine the hedging and safe haven abilities of energy and agricultural commodities against geopolitical risk. Their results suggest that, prior to and during the COVID-19 outbreak, commodities offer strong safe haven traits against uncertainties.

Mujtaba et al. [74] investigate the hedging and safe haven attributes of energy and agricultural commodities for the broader and sector-specific equity markets of both the United States and China. Their results suggest that agricultural commodities exhibit safe haven characteristics across all equity sectors in both countries. Conversely, energy commodities demonstrate safe haven qualities solely within the information technology and healthcare sectors. Bunditsakulporn [75] investigates the diversification, hedging, and safe haven abilities of agricultural commodities for the Thai stock market. The findings indicate that agricultural commodities, such as wheat, possess significant safe haven attributes in the Thai stock market. Furthermore, incorporating these particular agricultural commodities (either as a safe haven or a hedge) into a portfolio of Thai stocks is suggested to reduce risk and enhance performance in both standard and extreme downturn scenarios.

Ali et al. [55] examine the diversifying, hedging, and safe haven abilities of commodities against numerous global stock markets, suggesting that they possess valuable hedging and safe haven potentials. Using quantile-based approaches, Yang et al. [76] state that crude oil acts as a diversifier against shocks related to stocks. Tarchella, Khalfaoui, and Hammoudeh [77] examine the diversifying, hedging, and safe haven ability of crude oil against the G7 stock market indices during the COVID-19 pandemic. Their results support oil’s hedging ability during such crises. Utilizing bivariate and multivariate wavelet analyses, Woode, Idun, and Kawor [78] investigate the correlation between agricultural commodities and sub-Saharan African equities, which are significantly impacted by shocks from the COVID-19 pandemic. Their findings unveil the safe haven and hedging capabilities of commodities during both pandemic-induced and normal periods. In addition, their results indicate that international investors may contemplate investing in agricultural commodities and the selected equities to enhance portfolio diversification.

3. Data and Descriptive Statistics

3.1. Data

Our study period ranges from January 4, 2016, to July 5, 2023, encompassing black swan events, namely, the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. This period is divided into four distinct phases: a period of stability from January 4, 2016, to December 30, 2019; the COVID-19 era from December 31, 2019 (the onset of the pandemic), to February 23, 2022; the Russia–Ukraine war from February 24, 2022 (the start of the conflict), to March 9, 2023; and the SVB collapse from March 10, 2023 (the first day of the collapse), to July 5, 2023. It includes 1935 daily observations of three main energy and nonenergy commodities: crude oil (WTI), natural gas (NGK2), and wheat (ZWK2), along with the G7 stock market indices (United States [S & P 500], the United Kingdom [FTSE], Japan [NIKKEI], France [CAC 40], Germany [DAX 40], Italy [FTSE-MIB], and Canada [S & P/TSX]) and the G7 banking sector stock indices (USA, United Kingdom, Japan, Canada, France, Germany, and Italy). Data regarding commodities, the G7 stock market indices, and the G7 banking sector stock indices were obtained from the DataStream database. All price sequences are converted into natural logarithms and are characterized by $\begin{matrix} (1) & r_{it} = \log \frac{p_{i, t}}{p_{i, t - 1}}, \end{matrix}$ where $p_{i, t}$ ( $p_{i, t - 1}$ ) represents the closing price of asset i at time t ( $t - 1$ ).

3.2. Descriptive Statistics

The descriptive statistics summary of the return series before and during the COVID-19 pandemic, the Russia–Ukraine conflict, and the SVB collapse is presented in Table 1. It shows that before the COVID-19 outbreak, all the G7 stock market indices, the two commodities (WTI and wheat), and most of the G7 banking sector stock exhibit positive mean values. Crude oil takes the highest mean return while natural gas recorded a negative mean return. Natural gas is considered the highest volatile asset given its standard deviation value (0.012), followed by WTI (0.009), whereas, wheat is considered the least volatile asset. The skewness coefficients indicate that, before the COVID-19 outbreak, most of the assets displayed a leftward skew, as they possess negative values, except for WTI, wheat, and the Japanese banking sector index, which exhibit positive skewness. Substantial leptokurtosis is evident in all the return sequences and the Jarque–Bera test rejects the null hypothesis that the return series follows a normal distribution.

Table 1

Descriptive statistics.

Before COVID-19
Variable	Observations	Mean	Std. dev.	Skewness	Kurtosis	Jarque–Bera
Commodities
WTI	1027	0.000228	0.009511	0.362021	7.061276	728.2371
Wheat	1027	0.000083	0.007408	0.268735	3.930387	49.40274
Natural gas	1027	−0.000036	0.012251	−0.103711	8.148264	1136.018
G7 stock market indices
S & P 500	1027	0.000199	0.003481	−0.639041	7.832790	1069.336
FTSE	1027	0.000096	0.003460	−0.140890	5.364356	242.6107
NIKKEI	1027	0.000108	0.005008	−0.453817	10.54388	2470.529
DAX 40	1027	0.000108	0.004227	−0.590258	6.971928	734.7256
CAC 40	1027	0.000120	0.004139	−0.852464	10.77620	2711.966
FTSE-MIB	1027	0.000054	0.005594	−1.128651	15.54737	6955.007
S & P/TSX	1027	0.000118	0.002581	−0.359416	5.359280	260.2983
G7 banking sector stock indices
USA	1027	0.000208	0.004550	−0.469030	6.259493	492.2859
Canada	1027	0.000099	0.002934	−0.810378	9.025301	1665.927
UK	1027	0.000015	0.005397	−0.554928	10.67609	2574.094
France	1027	0.000039	0.004918	−1.308509	19.74382	12,289.95
Germany	1027	−0.000277	0.009309	−0.145554	6.830868	631.6175
Italy	1027	−0.000184	0.009574	−1.147914	20.93376	13,988.19
Japan	1027	−0.000207	0.006701	0.038444	7.122875	727.6300

During COVID-19
Variable	Observations	Mean	Std. dev.	Skewness	Kurtosis	Jarque–Bera

Commodities
WTI	554	0.000313	0.021517	−3.129554	52.05745	56,661.27
Wheat	554	0.000311	0.007455	0.353653	3.168849	12.20624
Natural gas	554	0.000585	0.017529	0.266209	5.303362	129.4772
G7 stock market indices
S & P 500	554	0.000212	0.006939	−1.015531	18.34871	5553.235
FTSE	554	−0.000092	0.006009	−1.233180	16.62575	4442.070
NIKKEI	554	0.000077	0.005927	0.101564	7.337599	436.8313
DAX 40	554	0.000076	0.006824	−1.031450	16.89622	4572.184
CAC 40	554	0.000098	0.006768	−1.365426	16.46704	4374.302
FTSE-MIB	554	0.000077	0.007410	−2.892952	33.27301	22,006.75
S & P/TSX	554	0.000151	0.006498	−1.796710	31.98113	19,756.96
G7 banking sector stock indices
USA	554	0.000174	0.009231	−0.629460	15.01516	3381.151
Canada	554	0.000208	0.007609	−0.600274	28.25409	14,808.37
UK	554	−0.000064	0.009094	−0.126821	8.070754	597.1644
France	554	0.000033	0.008797	−0.964262	14.94588	3392.137
Germany	554	0.000379	0.011817	−0.817136	11.20943	1623.187
Italy	554	0.000122	0.009374	−1.565748	17.15048	4865.979
Japan	554	0.000061	0.006504	−0.267047	5.378926	137.7153

During the Russia–Ukraine war
Variable	Observations	Mean	Std. dev.	Skewness	Kurtosis	Jarque–Bera

Commodities
WTI	270	−0.000315	0.012920	−0.436377	4.191144	24.53088
Wheat	270	−0.000441	0.013447	0.819590	9.885621	563.6102
Natural gas	270	−0.000961	0.023374	−0.443773	3.317700	9.997559
G7 stock market indices
S & P 500	270	−0.000121	0.006309	−0.083441	3.652809	5.107602
FTSE	270	0.000080	0.004187	−0.369187	6.063523	111.7167
NIKKEI	270	0.000136	0.005036	0.060862	3.700265	5.683363
DAX 40	270	0.000107	0.006046	0.355597	6.476895	141.6892
CAC 40	270	0.000122	0.005796	0.254206	6.262451	122.6483
FTSE-MIB	270	0.000105	0.006441	−0.396536	6.063354	112.6474
S & P/TSX	270	−0.000052	0.004230	−0.132677	3.745424	7.043281
G7 banking sector stock indices
USA	270	−0.000218	0.006480	0.165355	3.602272	5.311146
Canada	270	−0.000191	0.004330	−0.099922	3.626200	4.860724
UK	270	0.000131	0.007271	−0.348957	5.370536	68.69840
France	270	0.000039	0.006771	−0.154199	6.338120	126.4292
Germany	270	0.000065	0.011721	−0.779100	6.965502	204.2234
Italy	270	0.000165	0.010238	−0.647624	7.561710	252.9773
Japan	270	0.000382	0.005654	−0.041152	4.331455	20.01990

During the SVB collapse
Variable	Observations	Mean	Std. dev.	Skewness	Kurtosis	Jarque–Bera

Commodities
WTI	84	−0.000276	0.010529	−0.358780	3.353469	2.239416
Wheat	84	0.000066	0.009571	0.430516	3.658263	4.111405
Natural gas	84	0.000227	0.019189	0.071097	2.251412	2.032111
G7 stock market indices
S & P 500	84	0.000654	0.003455	0.178517	2.732924	0.695810
FTSE	84	−0.000296	0.155459	−0.046857	41.95637	5311.627
NIKKEI	84	0.000788	0.004139	−0.211134	2.765317	0.816849
DAX 40	84	0.00010	0.003987	−0.898918	5.274741	29.42332
CAC 40	84	−0.000036	0.004223	−0.799980	4.911049	21.74193
FTSE-MIB	84	0.000094	0.005277	−0.939979	6.060634	45.15604
S & P/TSX	84	0.000044	0.002879	−0.150316	3.091977	0.345938
G7 banking sector stock indices
USA	84	0.000062	0.008678	−0.693731	17.88459	782.1663
Canada	84	−0.000133	0.003830	−0.465373	3.082591	3.055877
UK	84	−0.000307	0.006582	−1.221021	5.832213	48.94752
France	84	−0.000497	0.006550	−1.456930	7.792495	110.1051
Germany	84	−0.000746	0.010904	−0.856729	5.855202	38.80843
Italy	84	−0.000035	0.009027	−1.082418	6.330666	55.22946
Japan	84	0.000059	0.006306	−1.351554	6.499100	68.42671

During the COVID-19 pandemic, all assets recorded positive mean values except for the FTSE stock index and the UK banking stock. During this health crisis, WTI and natural gas are considered the highest volatile assets. All assets display a leftward skew during this health crisis (except for wheat, natural gas, and NIKKEI) and exhibit significant leptokurtosis. Also, the results of the Jarque–Bera test reject the null hypothesis that the return series follows a normal distribution.

Considering the period of the Russia–Ukraine war, the three commodities along with the two G7 stocks (S & P 500 and S & P/TSX) and the two banking sector indices (USA and Canada) recorded negative mean values. In fact, the three commodities are considered the most volatile assets during this conflict. This may be explained by the importance of Russia and Ukraine as major exporters of crude oil, gas, and wheat to the rest of the world [23, 53, 79]. Moreover, most of the G7 stock market indices and banking sector stock indices recorded high volatility during this conflict. This may be justified by the engagement of the G7 stock market with the Russian market [26, 27]. All assets display a leftward skew during this crisis (except for wheat, NIKKEI, DAX 40, CAC 40, and the American banking stock) and exhibit significant leptokurtosis. Also, the results of the Jarque–Bera test reject the null hypothesis that the return series follows a normal distribution.

Regarding the period of the SVB collapse, WTI, the two G7 stock market indices (FTSE and CAC 40), and all the G7 banking sector stock indices (except for the US and Japan) recorded negative mean values supported by high volatility. All the G7 banking sector stock indices, the G7 stock market indices (except for the S & P 500), and WTI are left skewed. As for the rest of the assets, they are skewed to the right. All assets exhibit significant leptokurtosis values during this crisis and the results of the Jarque–Bera test reject the null hypothesis that the return series follows a normal distribution.

4. Methodology

4.1. Wavelet Coherence Approach

Following the definitions of Baur and Lucey [35], we examine the diversifying, hedging, and safe haven abilities of energy and nonenergy commodities (crude oil, natural gas, and wheat) against the G7 stock market indices and banking sector indices during the COVID-19 pandemic, the ongoing Russian–Ukrainian war, and the SVB collapse. To do so, we applied the wavelet coherence approach of Torrence and Compo [80] to examine the correlation between these assets and to identify their abilities as diversifying, hedging, and safe havens.

In fact, the methodology employed by Baur and Lucey [35] is primarily rooted in its reliance on linear correlation measures, which may not fully capture the dynamic and multiscale relationships between asset prices and market indices. This approach can overlook important temporal patterns and the complex interactions that evolve over different time horizons. However, the wavelet coherence approach addresses these limitations by allowing for the examination of both the time and frequency domains simultaneously. This method provides a more nuanced and comprehensive analysis of the comovement between commodities and stock market indices, revealing how their relationships vary across different time scales and during specific periods of market stress. By capturing these intricate details, the wavelet coherence approach offers a deeper understanding of the hedging and safe haven properties of crude oil, natural gas, and wheat, ultimately leading to more robust and insightful findings compared to the traditional methods used by Baur and Lucey [35].

We begin by computing the corresponding combinations of continuous wavelet transform (CWT), resulting in the derived cross-wavelet transform (XWT) and wavelet transform coherence (WTC). The CWT is excellent at capturing both the temporal and frequency aspects of time series data. The XWT enables the identification of areas within the time-frequency spectrum where the examined variables exhibit significant shared power and their corresponding phase alignment, whereas the WTC provides insights into the localized correlation between the two time series.

Undoubtedly, the wavelet coherence method plays a significant role in explaining the relationship between commodities and the G7 stock market indices, as well as the G7 banking sector indices, across diverse frequencies. This holds significance as financial markets encompass a multitude of participants with differing temporal outlooks, including short-term traders and long-term stakeholders. The correlation between commodities and indices relies on the frequencies at which these participants engage. Wavelet examination aids in recognizing slow and consistent movements, offering a more comprehensive understanding of market interconnectedness compared to conventional approaches. It identifies the optimal time intervals for the examined assets, unveiling whether they function as hedging tools, diversifiers, or safe haven options for financial markets, demonstrating either the absence of correlation or an inverse one across various time spans [81]. What sets the wavelet coherence method apart from conventional models is its capability to capture the comovement between two time series across both time and frequency domains. This strategy depends on a bivariate structure utilizing a wavelet transformation (Morlet set at 6), facilitating the acquisition of various scale positions [82]. To comprehend and interpret the comovement between time series across both temporal and frequency spectrums, we advocate for the wavelet coherence method, employing both XWT and coherence.

4.2. CWT Approach

This current manuscript utilizes the CWT as our primary instrument. Drawing upon the discoveries of Percival and Walden [83] and Aguiar-Conraria, Azevedo, and Soares [84], we can generate a set of wavelets for a given time-varying signal, $x t$ , through the process of enlarging and relocating a designated function, φ, recognized as the parent wavelet. Hence, we can affirm that the wavelet $φ_{μ, s} t$ is a square-integrable function in the real domain, characterized by its scale s, position μ, and temporal variable t, which can be specified as $\begin{matrix} (2) & φ_{μ, s} t = \frac{φ t - μ / s}{\sqrt{s}}, with φ . \in L^{2} r and S \neq 0 . \end{matrix}$

In this context, μ and s denote the temporal position and scale, correspondingly. $L^{2} r$ denotes the Hilbert space of one-dimensional functions that are square integrable, while $1 / \sqrt{s}$ signifies a coefficient that standardizes the wavelets to ensure they possess uniform variance (specifically, ${φ_{μ, s}}^{2} = 1$ ). This element is crucial to guarantee the comparability of various wavelet transformations across different scales and temporal positions [82, 85]. Therefore, we can reconstruct any temporal sequence from its wavelet conversions, given that the following condition is satisfied, wherein φ (f) is considered acceptable. $\begin{matrix} (3) & C_{φ} \int_{0}^{+ 0} \frac{{φ f}^{2}}{f}, with df < + \infty . \end{matrix}$

Based on our study, we utilize the Morlet¹ wavelet to investigate the wavelet coherence between energy and nonenergy commodities and the G7 stock market indices and banking sector indices. As per Grinsted, Moore, and Jevrejeva [87], the Fourier period of the Morlet wavelet is nearly identical to the scale applied. $\begin{matrix} (4) & φ_{ω_{0}} t = \frac{1}{\sqrt[4]{π}} e^{i ω_{0} t} e^{{- t}^{2} / 2} . \end{matrix}$

In this equation, and according to Torrence and Compo [80], t denotes the dimensionless time. The factor $\sqrt[4]{π}$ is a normalization factor ensuring that the wavelet possesses consistent energy. The exponent $e^{i ω_{0} t}$ denotes a complex sinusoidal component (sine wave), while the exponent $e^{{- t}^{2} / 2}$ represents a Gaussian-shaped envelope with a standard deviation of one unit, and $ω_{0}$ signifies the central angular frequency of the wavelet, dictating the frequency of oscillations within it. In our study, we adopt $ω_{0} = 6$ as the central frequency for the Morlet wavelet, as it offers optimal localization between time and frequency [80, 84, 87]. As $ω_{0}$ increases, frequency precision improves, while time precision diminishes, and vice versa. Following previous research (e.g., [88, 89]), we establish $ω_{0} = 6$ , fulfilling the admissibility criterion of the wavelet transformation.

We can classify wavelet transformations using various approaches. Initially, from the perspective of orthogonality, nonorthogonal wavelets are advantageous for CWT. Conversely, we employ orthogonal wavelets for the development of discrete wavelet transform (DWT). Second, as outlined by In and Kim [90], the continuous transformation of a signal relies on the scale and position parameters, each associated with a wavelet coefficient. Due to DWT, a specific signal is divided into orthogonal wavelets, and the frequency is roughly depicted on a logarithmic scale, followed by a portrayal of the displacement parameter, which is connected to the frequency [91]. We employ coherence and phase-difference techniques based on CWT [87], as we are interested in revealing localized correlations and temporal precedence relationships.

We can present the CWT $W_{x}$ of a discrete temporal sequence ( $x t$ , t = 0, 1, …, n) concerning $φ t$ in the subsequent manner. $\begin{matrix} (5) & W_{x} μ, s = \int_{- \infty}^{+ \infty} x t φ_{μ, s}^{'} t = \int_{- \infty}^{+ \infty} \frac{x t φ^{'} t - μ / s}{\sqrt{s}}, \end{matrix}$ where $φ^{'} \cdot$ represents the complex conjugate of $φ \cdot$ . Remarkably, the wavelet transformation conserves the energy of a temporal sequence, which can be employed to examine the power spectra. Consequently, the variance is expressed as $\begin{matrix} (6) & x^{2} = \frac{1}{C_{φ}} \int_{0}^{+ \infty} \int_{- \infty}^{+ \infty} W_{x} {μ, s}^{2} d μ \frac{d s}{s^{2}} . \end{matrix}$

4.3. XWT Approach

In addition to its utility in analyzing univariate time series data, wavelet analysis techniques can be extended to cover bivariate and multivariate scenarios. These applications involve addressing substantive concerns such as covariation patterns and causal links among variables across different scales and temporal spans. We employ the wavelet coherence approach to assess the degree of synchrony between two series of moments. Torrence and Compo [80] describe the XWT of two time series $x t$ and $y t$ using their respective XWTs (CWT) $W_{x} μ, s$ and $W_{y} μ, s$ as follows: $\begin{matrix} (7) & W_{x y} μ, s = W_{x} μ, s . W_{y} μ, s . \end{matrix}$

Due to the intricacy of the XWT, we frequently utilize the cross-wavelet power $W_{x y} μ, s$ to gauge the connection between two distinct time series. This enables us to identify the regions in the time-frequency domain where both series exhibit significant power; this phenomenon is referred to as their localized covariance. However, due to the unbounded nature of CWT outcomes, they are considered constrained. To address this concern, we employ the wavelet coherence $R_{xy}^{2} μ, s$ along with a smoothing operator S. Their calculations are as follows: $\begin{matrix} (8) & R_{xy}^{2} S = \frac{{S 1 / s W_{x y} μ, s}^{2}}{S 1 / s {W_{x} μ, s}^{2} S 1 / s {W_{y} μ, s}^{2}}, \end{matrix}$ where $R^{2}$ signifies the wavelet squared coherence, while $S \cdot$ denotes the smoothing operator. The squared coherence of the wavelet ranges from zero (indicating no comovement) to one (suggesting high comovement). This metric is regarded as a scale-dependent squared correlation between the series. In addition, the wavelet coherence framework allows us to examine the lead-lag dynamics between the series. This approach helps us address the issue of squared coherence, which hinders our ability to differentiate between positive and negative relationships among the series. Torrence and Compo [80] outline the definition of the phase difference in wavelet coherence as follows: $\begin{matrix} (9) & ψ_{x y} μ, s = \tan^{- 1} \frac{Φ S 1 / s W_{x y} μ, s}{Ω S 1 / s W_{x y} μ, s}, \end{matrix}$ where Φ and Ω represent the corresponding imaginary and real components of the smoothed XWT. The graphical outcome of the function is illustrated as an arrow. When the arrow points to the west (east) on the graph, it signifies that the associated time series displays a negative (positive) correlation. If the arrow points southward (northward), it indicates that the first series leads the second series by an angle of $π / 2$ , and conversely. However, the result frequently combines attributes of both, featuring arrows pointing southwestward. This indicates a negative correlation, with the initial series having priority over the second.

5. Empirical Results and Discussion

Appendixes A–C (respectively, Appendixes D–F) represent the correlation between commodities (WTI, gas, and wheat) and the G7 stock market indices (respectively, the G7 banking sector stock indices) before and during the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. The horizontal axis indicates time, whereas the vertical axis indicates the duration in days. In a color-coded display, we showcase the squared wavelet coherence value alongside its corresponding relative phase through arrows. In the wavelet coherence figures, colors represent the strength of correlation, where yellow represents high coherence and blue represents low coherence. A color scale ranging from 0 to 1 is displayed on the right-hand side of the graphs. The arrows are indicators of the nature of the correlation. When they point eastward, it indicates a positive correlation between the two series, suggesting a diversifying impact. Conversely, westward-pointing arrows suggest a negative correlation, meaning a hedge during regular periods and a safe haven during crises [35, 36]. When the arrows point upward, it indicates that the second series leads the first. Conversely, if the arrows point downward, it suggests that the first series guides the second. Ultimately, arrows positioned at the bottom indicate the direction of the long-term relationship.

5.1. Wavelet Coherence Between Commodities and the G7 Stock Market Indices

Appendix A depicts the correlation between WTI and the G7 stock market indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A1, A2, A3, A4, A5, A6, and A7 of this appendix, we can see that, before the COVID-19 outbreak, there exist big islands of coherency that show a strong positive correlation between these assets, with arrows pointing northeast (WTI taking the lead) or southeast (G7 stocks taking the lead). This positive correlation implies the diversifying ability of crude oil against the G7 stock market indices during times of stability. Before the COVID-19 outbreak, we can also observe small islands of negative coherency with arrows pointing west. This situation is observed only for the Japanese, German, and Italian stock market indices, suggesting the hedging ability of crude oil for these assets in the short run during times of stability. This supports the findings of Basher and Sadorsky [69] and Tarchella, Khalfaoui, and Hammoudeh [77] on the hedging ability of crude oil.

Moreover, the diversifying ability of WTI continues during the COVID-19 pandemic and the Russia–Ukraine military conflict for all the G7 stock market indices as evidenced by the big islands of strong positive coherency detected in all the figures with the arrows pointing east. This finding is in line with Ghorbel, Frikha, and Manzli [11], who state that WTI serves as a diversifying asset for the G7 stock market indices during the COVID-19 pandemic. It is also in line with Yang et al. [76], who state that crude oil acts as a diversifier against shocks related to stocks. However, this situation reverses during the SVB collapse where we can detect small islands of negative correlation with the arrows pointing mostly southwest and the G7 stocks taking the lead, suggesting the safe haven ability of crude oil during this financial crisis. This finding corroborates those of Bouoiyour, Selmi, and Wohar [70]; Jeribi and Snene-Manzli [3]; and Diaconaşu, Mehdian, and Stoica [72] regarding the safe haven ability of crude oil during periods of crisis.

In the long run, the correlation between WTI and the G7 stock market indices was clear only for the case of the FTSE stock index where the arrows point northeast and WTI takes the lead, suggesting the diversifying ability of crude oil against the British stock index in the long run (see bottom of Figure A6). As for the rest of the G7 stocks, the correlation was not clear since we could not detect any arrows.

Appendix B depicts the correlation between natural gas and the G7 stock market indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A8, A9, A10, A11, A12, A13, and A14 of this appendix, we can see that, before the COVID-19 pandemic, there exist many islands of strong positive coherency between gas and the G7 stock market indices with the arrows pointing southeast and the G7 stocks taking the lead and many islands of strong negative coherency with arrows pointing northwest and natural gas taking the lead. This suggests the diversifying (positive correlation) and hedging (negative correlation) ability of natural gas against the G7 stock market indices in the short run during periods of stability.

During the COVID-19 pandemic, natural gas acts as a strong safe haven for the French, German, British, and Italian stock market indices where we can detect small islands of negative coherency with the arrows pointing northwest and gas taking the lead. As for the American, Japanese, and Canadian stocks, gas acts as a weak safe haven. These findings are in line with Bandhu Majumder [51], who supports the safe haven ability of natural gas against stock markets during the COVID-19 pandemic.

During the Russia–Ukraine military conflict, natural gas acted mostly as a diversifier for the G7 stock market indices, except for the case of the French, German, and Italian stock indices for which it also showed safe haven capacities that were detected through small islands of negative correlation with arrows pointing northwest and gas taking the lead. This finding supports those of Kyriazis [54] regarding the safe haven ability of natural gas during the Russia–Ukraine conflict. As for the period of the SVB collapse, natural gas shows some safe haven patterns for all the G7 stock market indices. In the long run, the correlation between natural gas and the G7 stock market indices is not clear as we cannot detect any arrows (see the bottom of Figures A8, A9, A10, A11, A12, A13, and A14).

Appendix C illustrates the correlation between wheat and the G7 stock market indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A15, A16, A17, A18, A19, A20, and A21 of this appendix, we can see that, before the COVID-19 pandemic, wheat acts as a weak diversifier for all the G7 stock market indices in the short run as illustrated by the existence of small islands of correlation with the arrows pointing east. It also showed strong hedging patterns only for the French, German, and Italian stocks, which is illustrated by the existence of islands of negative correlations with the arrows pointing southwest.

During the COVID-19 pandemic, wheat acts mostly as a diversifier for all the G7 stock market indices as indicated by the arrows pointing mostly southeast. This corroborates the findings of Woode, Idun, and Kawor [78], who state that investors may consider investing in agricultural commodities to enhance stock portfolios’ diversification. This situation reversed during the Russia–Ukraine military conflict and the SVB collapse where we can observe islands of negative correlations with the arrows pointing northwest and wheat taking the lead, suggesting the safe haven ability of wheat against the G7 stocks in the short run during this political and financial crisis. These findings are in line with Bunditsakulporn [75]; Kyriazis [54]; Hasan et al. [52]; Mujtaba et al. [74]; and Woode, Idun, and Kawor [78], who support the safe haven ability of agricultural commodities, particularly wheat, during recent crises.

In the long run, Figures A15, A16, A17, A18, A19, A20, and A21 also show that wheat acts as a strong safe haven for all the G7 stock market indices (except for FTSE) as illustrated by the arrows pointing southwest.

To sum up, the analysis across Appendices A, B, and C reveals distinct correlations between commodities (WTI, natural gas, and wheat) and G7 stock market indices during various crises. Appendix A demonstrates WTI’s diversifying and hedging ability against G7 stocks pre-COVID-19, with a shift to safe haven status during the SVB collapse. Appendix B indicates gas’s positive and negative correlations pre-COVID-19, transitioning to a strong safe haven during the COVID-19 pandemic, particularly for certain European indices. In contrast, Appendix C highlights wheat’s weak diversifying role initially, evolving into a robust safe haven during crises, notably during the Russia–Ukraine conflict and the SVB collapse. Our findings underscore the safe haven role of energy and agricultural commodities against the G7 stocks during periods of crises, supporting the findings of Bouoiyour, Selmi, and Wohar [70]; Jeribi and Snene-Manzli [3]; Ji, Zhang, and Zhao [71]; Hasan et al. [52]; Kyriazis [54]; Bandhu Majumder [51]; Ghorbel, Frikha, and Manzli [11]; Bunditsakulporn [75]; Mujtaba et al. [74]; Tarchella, Khalfaoui, and Hammoudeh [77]; and Woode, Idun, and Kawor [78].

5.2. Wavelet Coherence Between Commodities and the G7 Banking Sector Stock Indices

Appendix D depicts the correlation between WTI and the G7 banking sector stock indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A22, A23, A24, A25, A26, A27, and A28 of this appendix, we can see that, before the COVID-19 outbreak, there are big islands of coherence that show a strong positive correlation between these assets in the short run, with arrows pointing northeast and WTI taking the lead. This positive correlation indicates the diversifying ability of crude oil against the G7 banking sector stock indices during times of stability. Before the COVID-19 outbreak, we can also observe small islands of negative coherence with arrows pointing west. This situation is observed only for the American, Japanese, and French banking indices, suggesting the hedging ability of crude oil for these assets in the short run during times of stability. This supports the findings of Basher and Sadorsky [69] and Tarchella, Khalfaoui, and Hammoudeh [77].

Moreover, the diversifying ability of WTI continues during the COVID-19 pandemic for all the G7 banking sector stock indices as evidenced by the big islands of strong positive coherence detected in all the Figures A22, A23, A24, A25, A26, A27, A28 with the arrows pointing east [11, 76]. However, this situation reversed during the Russia–Ukraine military conflict and the SVB collapse where we can detect many small islands of negative correlation with the arrows pointing mostly southwest and the G7 banking stocks taking the lead, suggesting the safe haven ability of crude oil during these geopolitical and financial crises [3, 70, 72, 78].

Nevertheless, WTI also shows some short-run diversifying abilities during the SVB collapse only for the French, German, and Italian banking stocks. In the long run, the correlation between WTI and the G7 banking sector stock indices was clear only for the case of the American, British, Canadian, and French banking stocks where the arrows point northeast and WTI takes the lead, suggesting the diversifying ability of crude oil against these banking stock indices in the long run (see bottom of Figures A22, A23, A25, and A26). As for the rest of the G7 banking sector stock indices, the correlation was not clear since we could not detect any arrows.

Appendix E depicts the correlation between natural gas and the G7 banking sector stock indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A29, A30, A31, A32, A33, A34, and A35 of this appendix, we can see that, before the COVID-19 pandemic, there are many small islands of either positive (arrows pointing east) or negative (arrows pointing west) coherence between gas and the G7 banking sector stock indices, suggesting the diversifying (positive correlation) and hedging (negative correlation) ability of gas against the G7 banking sector stock indices in the short run during periods of stability. During the COVID-19 pandemic, natural gas acts as a strong safe haven for all the G7 banking sector stock indices where we can detect small islands of negative coherence with the arrows pointing northwest and gas taking the lead. This supports the findings of Bandhu Majumder [51]. It also shows some diversifying abilities of the French, German, and Italian banking stocks during this health crisis. During the Russia–Ukraine military conflict and the SVB collapse, natural gas acts mostly as a strong safe haven for all the G7 banking stocks [54]. It also shows less pronounced diversifying abilities against these stocks during both the crises. In the long run, the correlation between natural gas and the G7 banking sector stock indices was clear only for the Italian bank stock index for which it acts as a safe haven as the arrows point northwest (see the bottom of Figure A35).

Appendix F illustrates the correlation between wheat and the G7 banking sector stock indices before and during the three crises of the COVID-19 pandemic, the Russia–Ukraine military conflict, and the SVB collapse. From Figures A36, A37, A38, A39, A40, A41, and A42 of this appendix, we can see that, before the COVID-19 pandemic, wheat acts as a diversifier for all the G7 banking sector stock indices in the short run as illustrated by the existence of many small islands of correlation with the arrows pointing east [78]. During the COVID-19 pandemic, wheat acts either as a safe haven or as a diversifier for all the G7 banking stocks as indicated by the arrows pointing west or east. This supports the findings of Woode, Idun, and Kawor [78], who highlight the diversifying and safe haven ability of agricultural commodities during crises. For the period of the Russia–Ukraine military conflict and the SVB collapse, wheat acts mostly as a short-run safe haven for all the G7 banking sector stocks as we can observe islands of negative correlations with the arrows pointing northwest and wheat taking the lead [52, 54, 74, 75, 78]. In the long run, Appendix F shows that wheat acts as a diversifier only for the American banking sector stock index and a strong safe haven for the Canadian, French, German, and Italian banking sector stock indices.

To sum up, the results of wavelet coherence between commodities and the G7 banking sector stock indices across Appendices D, E, and F reveal nuanced outcomes during different crises. As it is shown in Appendix D, WTI demonstrates a strong positive correlation before and during the COVID-19 pandemic, suggesting its diversifying ability, whereas, during the geopolitical and financial crises, it shows a negative correlation, suggesting its role as a safe haven asset. According to Appendix E, natural gas exhibited both positive and negative correlations before the pandemic, acting as both a diversifier and a hedge, but during crises, it primarily acts as a strong safe haven. Appendix F illustrates wheat’s role as a diversifier before the COVID-19 pandemic, with varied roles during this health crisis and other crises, acting as both a safe haven and a diversifier. In the long run, WTI correlates with a selection of banking stocks, indicating diversification potential, while natural gas shows a clear safe haven ability for the Italian banking sector, and wheat acts as a diversifier for the American banking stock and a strong safe haven for others. To conclude, our results validate the safe haven ability of energy and agricultural commodities against the G7 banking sector stock indices, corroborating the outcomes of Bouoiyour, Selmi, and Wohar [70]; Jeribi and Snene-Manzli [3]; Ji, Zhang, and Zhao [71]; Kyriazis [54]; Bandhu Majumder [51]; Hasan et al. [52]; Ghorbel, Frikha, and Manzli [11]; Bunditsakulporn [75]; Mujtaba et al. [74]; Tarchella, Khalfaoui, and Hammoudeh [77]; and Woode, Idun, and Kawor [78].

6. Conclusion

With the recent development of financial crises such as the COVID-19 pandemic, the Russian–Ukrainian conflict, and the SVB collapse, investors within financial markets often worry about numerous market-related uncertainties and risks, pushing them to seek suitable investment options that offer hedging and safe haven opportunities against potential threats from financial markets. In this context, our paper examines the hedging and safe haven abilities of energy commodities (crude oil and natural gas) and agricultural commodities (wheat) against both the G7 stock market indices and banking sector stock indices.

Our findings of wavelet coherence between commodities and G7 stock market indices provide important insights into their behavior throughout successive crises. Prior to the COVID-19 outbreak, commodities had varying correlations with the G7 stock markets, acting as both diversifiers and hedges. However, during crises such as the COVID-19 outbreak, the Russia–Ukraine conflict, and the SVB collapse, their roles shifted, with commodities frequently serving as powerful safe havens against market uncertainties. Specifically, WTI displayed diversification potential before COVID-19 but changed into a safe haven following the SVB crash. Natural gas mainly served as a safe haven during the pandemic, whereas wheat turned into a strong safe haven during crises. These findings underscore the importance of energy and agricultural commodities as safe haven assets during periods of market uncertainty and crises, aligning with prior research by Bouoiyour, Selmi, and Wohar [70]; Jeribi and Snene-Manzli [3]; Ji, Zhang, and Zhao [71]; Kyriazis [54]; Bandhu Majumder [51]; Hasan et al. [52]; Ghorbel, Frikha, and Manzli [11]; Bunditsakulporn [75]; Mujtaba et al. [74]; Tarchella, Khalfaoui, and Hammoudeh [77]; and Woode, Idun, and Kawor [78]. In addition, the analysis of wavelet coherence between commodities and the G7 banking sector stock indices further supports the notion of commodities as safe haven assets, providing valuable insights for investors navigating turbulent financial landscapes.

Our paper offers valuable insights into the safe haven potential of energy and agricultural commodities amid crises, expanding beyond traditional safe haven assets. It contributes to the understanding of portfolio diversification and risk management, providing investors with practical guidance for navigating volatile financial environments.

While this study provides valuable insights into the safe haven properties of crude oil, natural gas, and wheat against G7 stock market indices and banking sector stock indices during periods of crises, it is important to acknowledge its limitations. Specifically, the analysis was limited to the banking sector index due to its critical role and heightened sensitivity during economic stress, especially in light of events such as the SVB collapse. This focus was chosen to capture significant impacts on the financial system; however, it does not encompass the full spectrum of potential sectoral indices. Future research should consider expanding the scope to include other sectoral indices, such as technology, healthcare, and industrials, to provide a more comprehensive understanding of the hedging and safe haven capabilities of various commodities across different segments of the economy. This broader approach could offer more nuanced insights.

Funding

The study was supported by the Princess Nourah bint Abdulrahman University Researchers Supporting Project number: PNURSP2024R549, Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.

Endnotes

¹For further details regarding the Morlet wavelet, consult Addison [86].

Appendix

Appendix A: Wavelet Coherence Between Crude Oil (WTI) and the G7 Stock Market Indices