Abstract
Purpose
This paper aims to investigate whether the establishment of commodity futures can effectively hedge systemic risk in the spot network, given the context of financialization in the commodity futures market.
Design/methodology/approach
Utilizing industry association data from the Chinese commodity market, the authors identify systemically important commodities based on their importance in the production process using multiple graph analysis methods. Then the authors analyze the effect of listing futures on the systemic risk in the spot market with the staggered difference-in-differences (DID) method.
Findings
The findings suggest that futures contracts help reduce systemic risks in the underlying spot network. Systemic risk for a commodity will decrease by approximately 5.7% with the introduction of each corresponding futures contract, since the hedging function of futures reduces the timing behavior of firms in the spot market. Establishing futures contracts for upstream commodities lowers systemic risks for downstream commodities. Energy commodities, such as crude oil and coal, have higher systemic importance, with the energy sector dominating systemic importance, while some chemical commodities also have considerable systemic importance. Meanwhile, the shortest transmission path for risk propagation is composed of the energy industry, chemical industry, agriculture/metal industry and final products.
Originality/value
The paper provides the following policy insights: (1) The role of futures contracts is still positive, and future contracts should be established upstream and at more systemically important nodes in the spot production chain. (2) More attention should be paid to the chemical industry chain, as some chemical commodities are systemically important but do not have corresponding futures contracts. (3) The risk source of the commodity spot market network is the energy industry, and therefore, energy-related commodities should continue to be closely monitored.
Keywords
Citation
Liu, Q., Feng, Y. and Xu, M. (2023), "Will commodity futures reduce systemic risk in the spot market? Evidence from Chinese commodity market", China Finance Review International, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/CFRI-05-2023-0103
Publisher
:Emerald Publishing Limited
Copyright © 2023, Emerald Publishing Limited
1. Introduction
The futures markets are founded on two essential functions: risk management and price discovery. Although initially devised for risk hedging, the ongoing evolution of futures has transformed into sophisticated financial instruments. The process of financialization within these markets, and its resulting consequences, has garnered significant attention (Baker, 2021; Boyd et al., 2018; Natoli, 2021; Ordu-Akkaya and Soytas, 2020; Tang and Xiong, 2012; Zelazny, 2017). Notably, a discernible trend is the increased interconnectedness among commodities, as well as that between commodities and commodity indices or alternative asset classes (Natoli, 2021; Poncela et al., 2020). This trend can be attributed to the deliberate inclusion of commodities in the portfolios of non-industry investors, facilitating intricate linkages across diverse asset categories. As a result, these heightened interdependencies potentially disrupt asset valuations significantly and amplify the propagation of risks, thereby accentuating systemic vulnerabilities within the financial landscape. Illustratively, Engel (2020) meticulously examined the financialization of China's commodity market following the global financial crisis. Focusing on the Pan-Asian Metals Exchange and the market for commodity-backed financing, this study underscored that market opacity and unfounded expectations of perpetual asset price appreciation culminated in a surge of fraudulent activities. In the backdrop of financialization, the amplified interconnectedness of prices across various commodities accentuates systemic risk throughout the entire commodity network. Consequently, this study primarily investigates the effectiveness of commodity futures in mitigating systemic risk in the spot market, emphasizing their critical role in risk management.
While prior research has primarily emphasized the localized risk hedging effects of futures in the spot market, our study approaches this topic from a systemic risk perspective, employing a network framework to delve into the risk mitigation function of futures within the spot market, since the price volatility of commodities and futures not only determined by economic fundamentals but also vulnerable to the shock of other commodities due to financialization. Therefore, we mainly focus on the following issues: (1) the increasing financialization of commodity futures has amplified intricate interconnections among commodities, undeniably reinforcing their price interdependence. Is it crucial to identify and manage fluctuations in commodities with higher systemic importance to ensure the commodity market stability? (2) In the financialized commodity market, does the establishment of commodity futures still effectively help mitigate systemic risk in the spot market? (3)The close upstream and downstream production relationships among some commodities may form complex production networks. Will this physical network effect reinforce the connections within the spot market and thereby elevates systemic risk? If the establishment of futures can indeed continue to reduce systemic risk in the spot network, then which nodes are more suitable for introducing futures?
We choose to explore the above questions in the Chinese commodity futures market for two reasons. Firstly, China's commodity futures market is representative. According to the data from the Futures Industry Association (FIA), as of December 2021, the top ten agricultural futures and options contracts, as well as nine out of the top ten metal futures and options contracts traded globally are listed on China's three premier commodity futures exchanges, namely the Shanghai Futures Exchange, Dalian Commodity Exchange and Zhengzhou Commodity Exchange. In 2022, the trading volume within China's futures market surged with vigor, scaling a remarkable zenith at 53.49 trillion yuan. This resounding feat not only eclipsed the trading volume of stocks, valuated at 22.38 trillion yuan, but also outshone the bond market trading volume, commanding a robust 30.93 trillion yuan. Besides, the number of contracts has increased along with the growth in trading volume. A total of 65 commodities futures contracts, financial futures contracts, and 38 options contracts were listed in China's futures market as of the end of 2022. Secondly, as an emerging economy, China's commodity futures market has experienced a gradual maturation over the past three decades. In the context of financialization, the series of futures launched at different recent periods naturally provide an ideal test case for exploring the risk-constraint effect of commodity futures on the spot market. In the subsequent empirical section, we will employ a staggered DID model to investigate whether the introduction of futures effectively curbs systemic risk within the spot network.
Our paper is primarily related to three strands of literature. The first strand of literature focuses on the risk hedging capabilities inherent in futures markets. It has been found that the risk hedging function of futures has important impacts on enterprise value (Luo and Wang, 2018), operational risk (Bae et al., 2018; Chen et al., 2017), production decision-making (Zhao and Huchzermeier, 2017) and innovation (Blanco and Wehrheim, 2017). Moreover, scholars have extensively explored hedging strategies of futures, employing an array of statistical methodologies to ascertain optimal hedge ratios (Jeribi and Fakhfekh, 2021; Penone et al., 2021). Nevertheless, current literature has focused solely on localized risk hedging by futures, overlooking their potential systemic risk mitigation across the entire spot network, especially in the context of commodity futures financialization.
The second part of the literature discusses the price discovery function of futures. A substantial portion of this literature employs sophisticated statistical modeling techniques to scrutinize the intricate dance of returns and volatility spillovers, along with forecasting effects, between futures and the spot market. For instance, by using Value at Risk (VaR) or Vector Error Correction Model (VECM) models to test the cointegration and causal relationship between futures and spot prices (Bohl et al., 2011; Ge et al., 2019); examining the price leadership-lag relationship and dynamic changes between futures and spot prices (McCarthy and Orlov, 2012; Narayan and Sharma, 2018; Tonn et al., 2010); and using other methods such as quantile regression (Jena et al., 2019; Junior et al., 2020), nonlinear Granger causality test (Alzahrani et al., 2014), state-space model (Ge et al., 2019), dynamic conditional correlation (Rittler, 2012), to verify the price discovery function of commodity futures for spot prices. However, these studies predominantly narrow their gaze to the localized interplay between futures and spot prices, missing a comprehensive and systemic viewpoint. Despite their utilization of intricate statistical and econometric tools, these endeavors tend to stop short of a thorough exploration into the broader ramifications of commodity futures on the entirety of the spot market.
The third type of literature is the measurement of systemic risk and systemic importance. Many systemic risk measurements have been proposed in previous literature, among which the most popular ones are
Compared to the existing literature, the main contribution of this paper is twofold. Firstly, this study pioneers a systemic exploration of the risk-hedge role of futures on the spot market given the context of commodity futures financialization. More importantly, we pinpoint at which key nodes the establishment of futures is more conducive to hedging the systemic risk of the entire spot market. Our findings highlight that introducing futures at upstream nodes and at systemically important nodes in the spot network leads to a greater reduction of systemic risk, offering valuable insights for the future refinement of Chinese and global commodity futures market. Secondly, we apply an innovative approach by utilizing production-related data among various commodities. Leveraging the position and connectivity of commodities within production networks through a diversified graph analysis method, we identify systemically important commodities and the shortest risk transmission pathways among them. This innovation provides a more precise and directed reference for addressing systemic risk within the spot market.
The remaining parts of the paper will be unfolded as follows: Section 2 proposes the research hypotheses. Section 3 describes the econometric models and statistical methodologies employed in this article. Section 4 presents the empirical findings and finally the paper concludes.
2. Hypothesis
Risk management and price discovery are two major functions of the futures market. Due to their large trading volume, wide range of trading participants, transparent prices and fully digested information, futures prices have been an important variable affecting production-oriented enterprise decisions. Futures contracts can be utilized by enterprises to balance their operations in case of supply and demand mismatch in the spot market, which boost bargaining power, reduce corporate risks and expand financing channels (Bae et al., 2018; Chen et al., 2017; Luo and Wang, 2018; Zhao and Huchzermeier, 2017). However, as the financialization of futures markets continues, the correlation among commodity prices has become increasingly tight. The enhanced correlation could result in price fluctuations or risk contagion among assets (Engel, 2020; Natoli, 2021; Poncela et al., 2020), potentially affecting the spot market and escalating systemic risk. Hence, it is imperative to investigate if the introduction of commodity futures can still fulfill the function of managing risks and discovering prices and mitigating price fluctuations across the entire spot market, commonly known as systemic risk.
We argue that the establishment of futures will reduce the price volatility of the corresponding spot, since the hedging function of futures reduces the timing behavior of enterprises on the spot side. Futures can provide a channel for these companies to hedge their spot risk (Jeribi and Fakhfekh, 2021; Penone et al., 2021; Zhao et al., 2019), as most of the commodity spot traders are entity companies rather than financial investors. Suppose a company needs to purchase iron ore to produce metal products and now holds a certain amount of iron ore in spot. In the absence of futures, if the price of iron ore is expected to fall significantly in a relatively short period, the optimal behavior of the firm is to sell the iron ore and then buy it back when the price falls, a timing action that helps the firm reduce its purchasing costs. When all enterprises expect the spot price to fall and sell collectively, the selling behavior will further aggravate the spot price volatility and may even trigger the systemic risk of the whole spot market by positive feedback (Ahnert and Georg, 2018; Brunnermeier and Oehmke, 2014; Yarovaya et al., 2022). In contrast, with the establishment of the iron ore future, if firms expect spot prices to fall, they can hold short futures positions to hedge against spot risk, while the production of the spot will not be affected. Therefore, hypothesis 1 is proposed:
The establishment of corresponding commodity futures will reduce the systemic risk of commodities in the spot network.
Commodity markets form a complex production network based on industrial associations, which is a strong causal network. The fluctuations in commodity prices will propagate downstream according to the production relationships (Bargawi and Newman, 2017; Niu and Hu, 2021; Zhang et al., 2023). However, since production technology is basically unchanged in the short term, the production network is a static network structure. In order to investigate the dynamic changes of the commodity spot market network, according to Billio et al. (2012), Duan et al. (2021), Gong et al. (2019) and Yun et al. (2019), we construct a time-varying correlation network based on the causal relationship of spot returns. We hypothesize that the establishment of futures for upstream commodities will reduce their systemic risk, leading to a decrease in systemic risk transmitted downstream. Conversely, after the establishment of futures for downstream commodities, enterprises attempt to hedge risks of more associated commodities through the downstream nodes. This leads to an increase in upstream risk exposure and therefore increases the systemic risk of upstream commodities. Hence, we propose hypothesis 2:
The establishment of futures for upstream commodities will reduce the systemic risk of downstream commodities, while the establishment of futures for downstream commodities will increase the systemic risk of upstream commodities.
The introduction of futures contracts plays a crucial role in mitigating systemic risk associated with commodities. Going beyond traditional considerations of upstream and downstream relationships, we also examine the strategic positioning of commodities within the production network, a metric characterized by systemic importance (Acharya et al., 2017; Adrian and Brunnermeier, 2016; Banulescu and Dumitrescu, 2015; Bongini et al., 2015; Castro and Ferrari, 2014). Our objective is twofold: firstly, to identify commodities and sectors of heightened systemic significance within the production-centric spot network; and secondly, to delineate the shortest paths of risk transmission for each category based on production interconnections. This nuanced approach underscores that commodities with greater systemic importance attract increased investor motivation to hedge against price volatility risks. Consequently, the establishment of futures contracts results in a more pronounced reduction in systemic risk for such commodities, thereby enhancing overall market resilience. Then we propose hypothesis 3:
Spot commodities that are more systemically important in the production network will experience greater reductions in systemic risk after the establishment of futures.
3. Methodology
3.1 Directed graph
The commodity market is represented as a directed graph
3.1.1 Network of return relationships
Following Billio et al., (2012), Jeong and Park (2018) and Yun et al. (2019), we employ Granger causality testing to construct a network based on return relationships among commodity markets, rather than relying on the return correlations between commodities. The reason for this choice is that using correlations would result in an undirected graph, where the connections between commodity
3.1.2 Network of production relationships
The commodity market operates based on a complex supply chain system, with commodities being interdependent on one another. This supply chain relationship can be used to establish a network of production relationships among commodities. In this network, a directed causality relationship from commodity
The network of production relationships exhibits a “community” structure in which some nodes are clustered together based on their close production relationships. Figure 2 illustrates this community structure of the network of production relationships, where commodities are grouped into different “communities” according to the number and type of connections between nodes. Closely connected nodes are placed in the same community, which is characterized by a large number of internal connections and relatively few external connections. As shown in Figure 2, several chemical production chains form distinct community structures, represented by the two largest orange blocks, while chemical, energy and metal production chains form tightly knit communities, as shown in the largest blue block. This suggests that systemic risk may not only be transmitted among production networks of similar commodities, but also across different types of commodities.
Unlike the network of return relationships, we assume that the production technology remains stable in the short term, resulting in a strong and stable network of production relationships. The supply chain relationships in reality provide robust support for the construction of a directed graph, making the network of production relationships a more reliable measure of the systemic importance of commodities.
3.2 Network centrality
3.2.1 Structural holes
Structural holes refer to the lack of direct connections or ties between two or more entities in social networks (Burt, 2018). From the perspective of the network as a whole, it seems as though there are holes in the network structure due to the absence of direct connections or discontinuity between certain individuals. In the production network of commodity spot markets, if a commodity
3.2.2 PageRank
Historically, the PageRank algorithm was proposed as a way of computing the importance of web pages on the Internet (Page, 1998). The web linkage network is modeled as a directed graph, and a random walk model, specifically a first-order Markov chain, is defined on it to represent the process of a web surfer randomly browsing through web pages. Assuming that the web surfer visits each web page with equal probability and continuously jumps randomly on the network, this process is called a first-order Markov chain. PageRank represents the steady-state distribution of this Markov chain. The PageRank value of each web page is its steady-state probability. If web page
Therefore, when using the PageRank algorithm, the commodity spot market network we construct is different from that used by other algorithms. If commodity
The effect weight of each commodity in the network is normalized based on entity
After normalization, the PageRank algorithm is applied to obtain Rank.
The calculation of PageRank can be performed on the production network of commodity spot markets, usually through an iterative process. First, an initial distribution is assumed, and then all commodity PageRank values are calculated iteratively until convergence.
3.2.3 Degree centrality
In an undirected network, the centrality can be measured by the degree of a node. The underlying assumption behind this measure is that nodes with system importance are those that have many connections. The more connections a node has, the more important it is in the production network.
3.2.4 Closeness centrality
Degree centrality only utilizes the local feature of the network, which is how many connections a node has. However, having many connections does not necessarily mean that a commodity is located at the core of the network. Closeness centrality, similar to structural holes, utilizes the entire network feature, which is the position of a node in the entire structure. If a node has short distances to all other nodes in the network, its closeness centrality will be high. Compared to structural holes, closeness centrality is closer to the geometric center of the network.
3.2.5 Eigenvector centrality
The basic idea behind eigenvector centrality is that a node's centrality is a function of the centralities of its neighboring nodes. In other words, the more important the nodes connected to a node
Unlike degree centrality, where a node with many connections is considered central, a node with high eigenvector centrality may not have high degree centrality because its connections may have low eigenvector centrality. Similarly, a node with high eigenvector centrality does not necessarily have high degree centrality, as it may have few but important connections.
4. Data and empirical results
4.1 Data
This article uses web crawling technology to identify the industrial chain information of Chinese listed companies' annual reports and prospectuses and manually collects and constructs a production-based upstream and downstream network for 216 Chinese commodities. All industrial association data comes from disclosed files of listed companies, and the industry map is shown in Figure 1. At the same time, to ensure data consistency, this article collects weekly logarithmic returns of spot commodity prices from January 2012 to December 2022. The spot price data comes from the Wind database. Based on the spot return sequence, and referring to Billio et al. (2012), Gong et al. (2019) and Yun et al. (2019), a time-varying causal network of returns was constructed using Granger causality tests. Then, the PageRank value was calculated to measure the systemic risk of the spot, the number of connections for individual nodes (NodeCon) was calculated to measure the degree of connectivity between the commodity and other commodities, and the total number of connections in the entire spot network (NetClo) was calculated. We include these two indicators due to their inherent qualities. NodeCon provides insight into the degree of interconnectedness between individual nodes and their network counterparts, while NetClo offers a measure of the overall density of the network. Both metrics capture the intricate structure of the network's topology greatly, which is significant helpful in our analysis.
In addition, we controlled for a series of macroeconomic and financial variables at the weekly frequency, including international and domestic control variables such as the implied volatility of S&P 500 index (SP500VIX), the credit spread of US investment-grade bonds (CS_US), the logarithmic returns of the Reuters commodity index (dlnCRB), the implied volatility of the Shanghai Stock Exchange (SSE) 50 ETF (50ETF VIX), the credit spread of Chinese AA-grade bonds (CS_China) and the logarithmic returns of the Nanhua commodity index (dlnNanhua). The descriptive statistics of the variables are shown in Table 1. We collect futures listing dates for the sample period from Dalian Commodity Exchange, Zhengzhou Commodity Exchange and Shanghai Futures Exchange, as shown in Table 2.
4.2 Empirical results
4.2.1 Staggered introduction of corresponding futures
Different futures contracts have varying dates of listing, making the event of futures listing a multi-period exogenous policy. The prevailing literature focusing on the interplay between futures and spot markets has primarily employed statistical modeling techniques to delve into the first- and second-order moment effects, such as spillover effects, guiding relationships and predictive influences (Alzahrani et al., 2014; Bohl et al., 2011; Ge et al., 2019; Jena et al., 2019; Junior et al., 2020; McCarthy and Orlov, 2012; Narayan and Sharma, 2018). Nevertheless, we have not encountered research endeavor that regards the establishment of futures as an exogenous policy shock and scrutinizes its influence on the risk of spot market risk within the difference-in-differences (DID) framework. Thus, we apply a staggered DID model to examine whether the introduction of futures contracts diminishes the systemic risk associated with the particular commodity in the spot market network. The model for H1 is presented as follows:
In Columns (1)–(3), the coefficients of
This study employs a dynamic parallel trends test. Figure 3 shows that the treatment and control groups did not differ significantly in the first three years before the corresponding futures were introduced. However, in the year of and the third year after futures listing, both groups exhibit noteworthy differences. This suggests that the spot network's systemic risk dispersion meets the parallel trend assumption and that the effect of corresponding futures on reducing systemic risk in the spot market has persistence.
Moreover, a placebo test is conducted on the regression results by randomly assigning the listing time of all corresponding futures through 1000 Monte Carlo simulations. As shown in Figure 4, the placebo test does not observe any significant decrease in systemic risk in the spot market after corresponding futures are listed in the randomly assigned listing time groups. This test implies that the distribution of estimated parameters follows a normal distribution with a mean of 0. However, our true beta estimate in Table 2 is −0.00057, which is beyond the 95% confidential interval derived from the placebo test. These results suggest that the futures market's listing effect on reducing systemic risk in the spot market is robust.
4.2.2 Upstream and downstream corresponding futures
Then, to answer the question of which nodes in the spot market network are most beneficial to set up futures contracts to reduce systemic risk, this paper further investigates the impact of commodities from upstream and downstream on the concerned commodity after the establishment of futures contracts. The model for H2 is as follows:
Columns (1)–(3) in Table 4 control for a variety of variables, such as
4.2.3 Effects on systemic important commodity
Next, we examine whether there is heterogeneity in the effect of futures contracts on systemic risk for more systemically important commodities in the spot market network. Before that, we use graph analysis to identify the most systemically important commodities in the network.
Figure 5 illustrates the systemic importance network obtained from the structural holes, PageRank algorithms, point degree centrality, closeness centrality and eigenvector centrality, respectively. These graphs show that energy products, such as crude oil, coal, coke and methanol, are the most core and systemically important commodities in the commodity market, with several of them already having corresponding listed futures. Although sulfuric acid, trichloromethane and compound fertilizer are among the most systemically important chemical products in the commodity market, they have not received enough attention compared to energy products and corresponding listed futures or options contracts for these key production nodes are currently unavailable.
To obtain a clearer map of important commodity futures production relationships, this study further analyzed the sub-sample network that has corresponding listed futures in Figure 6. Energy products still occupy the central position as systemically important commodities in the sub-sample network. Additionally, urea is also a systemically important commodity in the commodity futures network because it can be used to produce agricultural fertilizers. The risk transmission path in the sub-sample network is similar to that of the entire sample, which supports the robustness of our results.
Table 5 presents the top ten commodities based on their structural holes and displays all their systemic importance indicators. Trichloromethane exhibits a centrality far exceeding other varieties, with a score of 3.151, indicating that it links multiple commodities that are not directly connected in the production chain and occupies a central position in the production chain, consistent with the actual production relationships. The results of PageRank, eigenvector centrality and point degree centrality are relatively similar, with trichloromethane, thermal coal and crude oil occupying important centralities. This suggests that these commodities not only serve as production materials for multiple commodities but also occupy an important position in the production relationships themselves and many commodities related to them also occupy important positions in the production relationships. However, the results of closeness centrality differ significantly from other indicators. In closeness centrality, methanol, coke, thermal coal and sulfuric acid occupy important central positions. This indicates that these bulk commodities have a relatively shorter minimum distance to other commodities in the production relationships, which is easy to understand. Although these commodities have slightly different systemic importance indicators than other commodities, they are also energy or chemical products that serve as direct raw materials for multiple commodities and occupy an important position in the production relationships.
To better understand the process of systemic risk transmission, this article presents a shortest path diagram based on the production network, as shown in Figure 7. The results show that the shortest path for risk transmission consists of the energy industry-chemical industry-agriculture/metals-final products. Combined with the results of Table 4, commodity futures should be established upstream in the production chain as much as possible. Since several energy commodities with the highest systemic importance have corresponding futures listed, a more sensible approach would be to focus on the chemical industry and establish futures contracts for its systemically important commodities, such as trichloromethane and sulfuric acid, which should effectively reduce the systemic risk of the entire spot market.
As previously analyzed, commodities with greater systemic importance have more direct relationships with other important commodities in the production chain. Therefore, it is important to explore whether the impact of futures trading on spot prices has heterogeneity in terms of systemic importance. The model for H3 is as follows:
In Columns (1)–(3) of Table 6, we control for a series of variables, including
In addition, according to Columns (2), if the upstream commodity of spot
In summary, our empirical findings substantiate that under financialization, commodity futures maintain a pronounced capacity to continuously reduce systemic risk within the spot network, which fills the gap of the existing literature. Besides, we reveal that the establishment of futures contracts at upstream nodes and at systemically important nodes within the spot network mitigates more systemic risk. This observation provides compelling evidence that guides the strategic evolution of China's commodity futures market. The concept of systemic importance has primarily been applied to systemic importance of financial institutions (Banulescu and Dumitrescu, 2015; Bongini et al., 2015; Castro and Ferrari, 2014), and there is currently scarce literature discussing systemic importance within commodity networks. Thus, our research pioneers the extension of systemic importance identification methodologies into the realm of commodity topology networks. Meanwhile, our innovatively use of production-linked data enhances the ability to identify the shortest pathways of risk transmission within the commodity topology network. This finding holds the potential to enable regulatory authorities to monitor and prevent systemic risk at their source.
5. Conclusion and policy insights
In the context of financialization in the commodity futures market, we investigate whether the establishment of commodity futures can still effectively hedge systemic risk in the spot network. Since heightened interdependencies caused by financialization have the potential to disrupt asset valuations significantly and amplify the propagation of risks, thereby accentuating systemic risk, our research offers considerable insights for preempting systemic risk within the spot market. It turns out that commodity futures are helpful to reduce systemic risk within the spot network, and systemic risk will decrease by approximately 5.7% with the introduction of each corresponding futures contract. Notably, our research underscores the strategic significance of siting futures contracts at upstream and systemically important nodes within the spot network, serving as a potent avenue for fortifying systemic risk mitigation. Simultaneously, we unveil the shortest risk transmission pathways within the spot network, which is energy industry - chemical industry - agriculture/metal industry - final products, shedding light on the origins and propagation process of systemic risk. In conclusion, our empirical findings resoundingly validate the continued effectiveness of commodity futures in adeptly mitigating systemic risk within the evolving financialization. This previously unexplored insight constitutes a substantive and compelling contribution to the prevailing academic discourse.
The paper provides the following policy insights: Firstly, the role of futures contracts is still positive, and future contracts should be established upstream and at more systemically important nodes in the spot production chain. Secondly, more attention should be paid to the chemical industry chain, as some chemical commodities are systemically important but do not have corresponding futures contracts. Thirdly, the risk source of the commodity spot market network is the energy industry, and therefore, energy-related commodities should continue to be closely monitored.
Figures
Figure 1
Network of production relationships
Figure 2
Community structure of the production relationship network
Figure 3
Dynamic parallel trend test
Figure 4
Placebo test
Figure 5
Systemic importance network obtained from difference measures based on production relationship network
Figure 6
Systemic importance network of commodities with listed futures obtained from difference measures based on production relationship network
Figure 7
The shortest risk transmission path
Summary statistics
| count | mean | std | min | 25% | 50% | 75% | max | |
|---|---|---|---|---|---|---|---|---|
| PageRank | 57,134 | 0.0102 | 0.0104 | 0.0030 | 0.0048 | 0.0066 | 0.0109 | 0.1385 |
| NodeCon | 57,134 | 17.05 | 8.46 | 0.00 | 11.00 | 16.00 | 22.00 | 62.00 |
| NetClo | 57,134 | 835.50 | 144.55 | 558.00 | 732.00 | 821.00 | 887.00 | 1325.00 |
| SP500 VIX | 55,370 | 17.540 | 6.883 | 9.140 | 12.920 | 15.650 | 20.600 | 66.040 |
| 50ETF VIX | 38,710 | 22.768 | 7.469 | 8.681 | 18.132 | 22.255 | 26.025 | 57.591 |
| CS_China (%) | 54,684 | 2.0887 | 0.6343 | 1.2292 | 1.5640 | 1.8248 | 2.6983 | 3.5402 |
| CS_US(%) | 55,370 | 1.7915 | 0.4711 | 1.0700 | 1.4700 | 1.7400 | 2.0300 | 4.6500 |
| dlnNanhua | 53,214 | 0.0008 | 0.0206 | −0.0928 | −0.0104 | 0.0003 | 0.0116 | 0.1088 |
| dlnCRB | 55,370 | −0.0002 | 0.0220 | −0.1298 | −0.0115 | −0.0003 | 0.0134 | 0.1256 |
Source(s): Authors' own work
Listed dates of Chinese commodity futures within the sample period
| Commodity | Listed date | Commodity | Listed date |
|---|---|---|---|
| Methanol | 2011/10/28 | Tin | 2015/3/27 |
| Silver | 2012/5/10 | Nickel | 2015/3/27 |
| Rapeseed | 2012/12/28 | Cotton yarn | 2017/8/18 |
| Glass | 2012/12/3 | Apples | 2017/12/22 |
| Rapeseed meal | 2012/12/28 | Crude oil | 2018/3/26 |
| Coking coal | 2013/3/22 | Viscose staple fiber | 2018/11/27 |
| Steam coal | 2013/9/26 | Ethylene glycol | 2018/12/10 |
| Asphalt | 2013/10/18 | Red dates | 2019/4/30 |
| Iron ore | 2013/10/18 | Japonica rice | 2019/8/16 |
| Eggs | 2013/11/8 | Urea | 2019/8/9 |
| Japonica rice | 2013/9/26 | Stainless-steel coil | 2019/9/25 |
| Fiberboard | 2019/12/2 | 20# Rubber | 2019/8/12 |
| Plywood | 2013/12/6 | Styrene | 2019/9/26 |
| Polypropylene | 2014/2/28 | Soda ash | 2019/12/6 |
| Ferro silicon | 2014/8/8 | Liquefied-petroleum gas | 2020/3/30 |
| Hot-rolled coil | 2014/3/21 | Low-sulfur fuel oil | 2020/6/22 |
| Silicon manganese | 2014/8/8 | Polyester short fiber | 2019/12/6 |
| Indica rice | 2014/7/8 | Live pigs | 2021/1/8 |
| Corn starch | 2014/12/19 | Peanuts | 2021/2/1 |
Source(s): Authors' own work
Panel regression for the impact of futures listing on spot network price transmission (unit: \%)
| (1) | (2) | (3) | |
|---|---|---|---|
| Commodity Characteristics | |||
| DIDList | −0.060*** | −0.063*** | −0.057*** |
| (−5.72) | (−5.99) | (−4.55) | |
| NodeCon | 0.029*** | 0.029*** | |
| (58.20) | (44.56) | ||
| NetClo | −0.001*** | −0.000*** | |
| (−15.86) | (−13.23) | ||
| Macroeconomic variables | |||
| SP500 VIX | 0.001 | ||
| (0.01) | |||
| CS_US | −0.001 | ||
| (−0.00) | |||
| CRB | 0.002 | ||
| (0.01) | |||
| SSE50 VIX | −0.004 | ||
| (−0.02) | |||
| CS_China | 0.002 | ||
| (0.01) | |||
| Nanhua | −0.006 | ||
| (−0.00) | |||
| Dummy variables | |||
| Industry FE | Control | Control | Control |
| Year FE | Control | Control | Control |
| Observations | 57,134 | 57,134 | 37,926 |
| Adj R-squared | 0.450 | 0.490 | 0.491 |
Note(s):
Source(s): Authors' own creation
Panel analysis of the impact of futures listing on the transmission of upstream and downstream spot prices (unit: \%)
| (1) | (2) | (3) | |
|---|---|---|---|
| Commodity characteristics | |||
| Upstream | −0.074*** | −0.076*** | |
| (−7.16) | (−7.31) | ||
| Downstream | 0.070*** | 0.072*** | |
| (−4.79) | (7.64) | ||
| DIDList | −0.057*** | −0.059*** | −0.059*** |
| (−4.53) | (−4.79) | (−4.77) | |
| NodeCon | 0.030*** | 0.028*** | 0.029*** |
| (44.37) | (41.97) | (42.00) | |
| NetClo | −0.006*** | −0.005*** | −0.006*** |
| (−13.48) | (−12.90) | (−13.15) | |
| Macroeconomic variables | |||
| SP500 VIX | 0.001 | −0.001 | −0.002 |
| (0.07) | (−0.08) | (−0.01) | |
| CS_US | −0.006 | 0.003 | −0.001 |
| (−0.01) | (0.01) | (−0.00) | |
| CRB | −0.011 | 0.003 | −0.009 |
| (−0.05) | (0.02) | (−0.05) | |
| SSE50 VIX | −0.002 | 0.001 | −0.000 |
| (−0.14) | (0.07) | (−0.05) | |
| CS_China | 0.005 | −0.002 | 0.002 |
| (0.17) | (−0.08) | (0.09) | |
| Nanhua | 0.007 | 0.001 | 0.009 |
| (0.04) | (0.01) | (0.05) | |
| Dummy variables | |||
| Industry FE | Control | Control | Control |
| Year FE | Control | Control | Control |
| Observations | 37,926 | 37,926 | 37,926 |
| Adj R-squared | 0.511 | 0.510 | 0.515 |
Note(s):
Source(s): Authors' own creation
Systemic importance measures of select commodities
| PageRank | Eigenvector Centrality | Closeness Centrality | Degree Centrality | Structural hole | |
|---|---|---|---|---|---|
| Crude oil | 0.038 | 0.181 | 0.268 | 0.031 | 0.826 |
| Coal | 0.031 | 0.287 | 0.315 | 0.037 | 1.590 |
| CHCL3 | 0.028 | 0.187 | 0.274 | 0.045 | 3.151 |
| Coking Coal | 0.025 | 0.078 | 0.266 | 0.008 | 0.132 |
| H2SO4 | 0.023 | 0.156 | 0.310 | 0.034 | 1.027 |
| MA | 0.021 | 0.233 | 0.318 | 0.028 | 1.645 |
| Coke | 0.021 | 0.253 | 0.319 | 0.025 | 1.215 |
| IO | 0.017 | 0.042 | 0.212 | 0.014 | 0.077 |
| LPG | 0.015 | 0.123 | 0.251 | 0.014 | 0.371 |
| NP | 0.014 | 0.163 | 0.293 | 0.014 | 0.692 |
Source(s): Authors' own creation
Heterogeneity analysis of futures listing impact on spot price transmission: systemic importance (unit: %)
| (1) | (2) | (3) | |
|---|---|---|---|
| Commodity characteristics | |||
| DIDList | −0.148*** | ||
| (−4.92) | |||
| DIDList | −0.025** | ||
| (−2.03) | |||
| Upstream | −0.040*** | ||
| (−4.99) | |||
| Upstream | −0.068*** | ||
| (−7.26) | |||
| Downstream | 0.012 | ||
| (1.51) | |||
| Downstream | 0.070*** | ||
| (7.39) | |||
| NodeCon | 0.029*** | 0.030*** | 0.028*** |
| (44.48) | (44.37) | (41.81) | |
| NetClo | −0.006*** | −0.006*** | −0.005*** |
| (−13.17) | (−13.46) | (−12.88) | |
| Macroeconomic variables | |||
| SP500 VIX | 0.000 | 0.005 | −0.002 |
| (0.00) | (0.23) | (−0.10) | |
| CS_US | 0.001 | −0.014 | 0.003 |
| (0.00) | (−0.25) | (0.06) | |
| CRB | 0.002 | 0.004 | −0.002 |
| (0.01) | (0.02) | (−0.01) | |
| SSE50 VIX | −0.002 | 0.000 | 0.002 |
| (−0.01) | (0.03) | (0.10) | |
| CS_China | 0.007 | 0.008 | −0.002 |
| (0.02) | (0.26) | (0.01) | |
| Nanhua | 0.002 | 0.010 | 0.001 |
| (0.00) | (0.06) | (0.01) | |
| Dummy variables | |||
| Industry FE | Yes | Yes | Yes |
| Year FE | Yes | Yes | Yes |
| Observations | 37,926 | 37,926 | 37,926 |
| Adj R-squared | 0.491 | 0.511 | 0.510 |
Note(s):
Source(s): Authors' own creation
References
Acharya, V.V., Pedersen, L.H., Philippon, T. and Richardson, M. (2017), “Measuring systemic risk”, The Review of Financial Studies, Vol. 30, pp. 2-47, doi: 10.1093/rfs/hhw088.
Adrian, T. and Brunnermeier, M.K. (2016), “CoVaR”, American Economic Review, Vol. 106, pp. 1705-1741, doi: 10.1257/aer.20120555.
Ahnert, T. and Georg, C.-P. (2018), “Information contagion and systemic risk”, Journal of Financial Stability, Vol. 35, pp. 159-171, doi: 10.1016/j.jfs.2017.05.009.
Alzahrani, M., Masih, M. and Al-Titi, O. (2014), “Linear and non-linear Granger causality between oil spot and futures prices: a wavelet based test”, Journal of International Money and Finance, Vol. 48, pp. 175-201, doi: 10.1016/j.jimonfin.2014.07.001.
Bae, S.C., Kim, H.S. and Kwon, T.H. (2018), “Currency derivatives for hedging: new evidence on determinants, firm risk, and performance”, Journal of Futures Markets, Vol. 38, pp. 446-467, doi: 10.1002/fut.21894.
Baker, S.D. (2021), “The financialization of storable commodities”, Management Science, Vol. 67, pp. 471-499.
Banulescu, G.-D. and Dumitrescu, E.-I. (2015), “Which are the SIFIs? A Component Expected Shortfall approach to systemic risk”, Journal of Banking and Finance, Vol. 50, pp. 575-588, doi: 10.1016/j.jbankfin.2014.01.037.
Bargawi, H.K. and Newman, S.A. (2017), “From futures markets to the farm gate: a study of price formation along Tanzania's coffee commodity chain”, Economic Geography, Vol. 93, pp. 162-184, doi: 10.1080/00130095.2016.1204894.
Bianchi, D., Billio, M., Casarin, R. and Guidolin, M. (2019), “Modeling systemic risk with Markov switching graphical SUR models”, Journal of Econometrics, Vol. 210, pp. 58-74, doi: 10.1016/j.jeconom.2018.11.005.
Billio, M., Getmansky, M., Lo, A.W. and Pelizzon, L. (2012), “Econometric measures of connectedness and systemic risk in the finance and insurance sectors”, Journal of Financial Economics, Vol. 104, pp. 535-559, doi: 10.1016/j.jfineco.2011.12.010.
Blanco, I. and Wehrheim, D. (2017), “The bright side of financial derivatives: options trading and firm innovation”, Journal of Financial Economics, Vol. 125, pp. 99-119, doi: 10.1016/j.jfineco.2017.04.004.
Bohl, M.T., Salm, C.A. and Schuppli, M. (2011), “Price discovery and investor structure in stock index futures”, Journal of Futures Markets, Vol. 31, pp. 282-306, doi: 10.1002/fut.20469.
Bongini, P., Nieri, L. and Pelagatti, M. (2015), “The importance of being systemically important financial institutions”, Journal of Banking and Finance, Vol. 50, pp. 562-574, doi: 10.1016/j.jbankfin.2014.07.006.
Boyd, N.E., Harris, J.H. and Li, B. (2018), “An update on speculation and financialization in commodity markets”, Journal of Commodity Markets, Vol. 10, pp. 91-104, doi: 10.1016/j.jcomm.2018.05.005.
Brownlees, C. and Engle, R.F. (2017), “SRISK: a conditional capital shortfall measure of systemic risk”, The Review of Financial Studies, Vol. 30, pp. 48-79, doi: 10.1093/rfs/hhw060.
Brunnermeier, M.K. and Oehmke, M. (2014), “Predatory short selling*”, Review of Finance, Vol. 18, pp. 2153-2195, doi: 10.1093/rof/rft043.
Burt, R.S. (2018), “Structural holes”, Social Stratification, Routledge, pp. 659-663.
Castro, C. and Ferrari, S. (2014), “Measuring and testing for the systemically important financial institutions”, Journal of Empirical Finance, Vol. 25, pp. 1-14, doi: 10.1016/j.jempfin.2013.10.009.
Chen, Z., Han, B. and Zeng, Y. (2017), “Financial hedging and firm performance: evidence from cross-border mergers and acquisitions: financial hedging and firm performance”, European Financial Management, Vol. 23, pp. 415-458, doi: 10.1111/eufm.12103.
Duan, Z., Gong, Z. and Qi, Q. (2021), “Factorization asset pricing”, SSRN Journal. doi: 10.2139/ssrn.3940074.
Engel, J. (2020), “Regulation, financialization and fraud in Chinese commodity markets after the Global Financial Crisis”, Regional Studies, Vol. 54, pp. 219-228, doi: 10.1080/00343404.2018.1537482.
Ge, Y., Cao, T., Jiang, R., Liu, P. and Xie, H. (2019), “Does China’s iron ore futures market have price discovery function? Analysis based on VECM and state-space perspective”, Journal of Business Economics and Management, Vol. 20, pp. 1083-1101, doi: 10.3846/jbem.2019.10604.
Gong, X.-L., Liu, X.-H., Xiong, X. and Zhang, W. (2019), “Financial systemic risk measurement based on causal network connectedness analysis”, International Review of Economics and Finance, Vol. 64, pp. 290-307, doi: 10.1016/j.iref.2019.07.004.
Härdle, W.K., Wang, W. and Yu, L. (2016), “TENET: tail-Event driven NETwork risk”, Journal of Econometrics, Vol. 192, pp. 499-513, doi: 10.1016/j.jeconom.2016.02.013.
Jena, S.K., Tiwari, A.K., Hammoudeh, S. and Roubaud, D. (2019), “Distributional predictability between commodity spot and futures: evidence from nonparametric causality-in-quantiles tests”, Energy Economics, Vol. 78, pp. 615-628, doi: 10.1016/j.eneco.2018.11.013.
Jeong, D. and Park, S. (2018), “The more connected, the better? Impact of connectedness on volatility and price discovery in the Korean financial sector”, MF, Vol. 44, pp. 46-73, doi: 10.1108/MF-09-2016-0277.
Jeribi, A. and Fakhfekh, M. (2021), “Portfolio management and dependence structure between cryptocurrencies and traditional assets: evidence from FIEGARCH-EVT-Copula”, J Asset Manag, Vol. 22, pp. 224-239, doi: 10.1057/s41260-021-00211-7.
Junior, P.O., Tiwari, A.K., Padhan, H. and Alagidede, I. (2020), “Analysis of EEMD-based quantile-in-quantile approach on spot- futures prices of energy and precious metals in India”, Resources Policy, Vol. 68, 101731, doi: 10.1016/j.resourpol.2020.101731.
Kanno, M. (2015), “Assessing systemic risk using interbank exposures in the global banking system”, Journal of Financial Stability, Vol. 20, pp. 105-130, doi: 10.1016/j.jfs.2015.08.006.
Kuzubaş, T.U., Ömercikoğlu, I. and Saltoğlu, B. (2014), “Network centrality measures and systemic risk: an application to the Turkish financial crisis”, Physica A: Statistical Mechanics and Its Applications, Vol. 405, pp. 203-215, doi: 10.1016/j.physa.2014.03.006.
Luo, H.(R.) and Wang, R. (2018), “Foreign currency risk hedging and firm value in China”, Journal of Multinational Financial Management, Vols 47-48, pp. 129-143, doi: 10.1016/j.mulfin.2018.11.002.
McCarthy, J. and Orlov, A.G. (2012), “Time-frequency analysis of crude oil and S&P500 futures contracts”, Quantitative Finance, Vol. 12, pp. 1893-1908, doi: 10.1080/14697688.2012.686669.
Narayan, P.K. and Sharma, S.S. (2018), “An analysis of time-varying commodity market price discovery”, International Review of Financial Analysis, Vol. 57, pp. 122-133, doi: 10.1016/j.irfa.2018.03.008.
Natoli, F. (2021), “Financialization of commodities before and after the great financial crisis”, Journal of Economic Surveys, Vol. 35, pp. 488-511, doi: 10.1111/joes.12408.
Niu, H. and Hu, Z. (2021), “Information transmission and entropy-based network between Chinese stock market and commodity futures market”, Resources Policy, Vol. 74, 102294, doi: 10.1016/j.resourpol.2021.102294.
Ordu-Akkaya, B.M. and Soytas, U. (2020), “Unconventional monetary policy and financialization of commodities”, The North American Journal of Economics and Finance, Vol. 51, 100902, doi: 10.1016/j.najef.2018.12.014.
Page, L. (1998), The Pagerank Citation Ranking : Bringing Order to the Web. Technical Report, Stanford Digital Library Technologies Project, 1998.
Penone, C., Giampietri, E. and Trestini, S. (2021), “Hedging effectiveness of commodity futures contracts to minimize price risk: empirical evidence from the Italian field crop sector”, Risks, Vol. 9, p. 213, doi: 10.3390/risks9120213.
Poncela, P., Senra, E. and Sierra, L.P. (2020), “Global vs sectoral factors and the impact of the financialization in commodity price changes”, Open Econ Rev, Vol. 31, pp. 859-879, doi: 10.1007/s11079-019-09564-4.
Rittler, D. (2012), “Price discovery and volatility spillovers in the European Union emissions trading scheme: a high-frequency analysis”, Journal of Banking and Finance, Vol. 36, pp. 774-785, doi: 10.1016/j.jbankfin.2011.09.009.
Tang, K. and Xiong, W. (2012), “Index investment and the financialization of commodities”, Financial Analysts Journal, Vol. 68, pp. 54-74, doi: 10.2469/faj.v68.n6.5.
Tonn, V.L., Li, H.C. and McCarthy, J. (2010), “Wavelet domain correlation between the futures prices of natural gas and oil”, The Quarterly Review of Economics and Finance, Vol. 50, pp. 408-414, doi: 10.1016/j.qref.2010.07.002.
Yarovaya, L., Brzeszczyński, J., Goodell, J.W., Lucey, B. and Lau, C.K.M. (2022), “Rethinking financial contagion: information transmission mechanism during the COVID-19 pandemic”, Journal of International Financial Markets, Institutions and Money, Vol. 79, 101589, doi: 10.1016/j.intfin.2022.101589.
Yun, T.-S., Jeong, D. and Park, S. (2019), “‘Too central to fail’ systemic risk measure using PageRank algorithm”, Journal of Economic Behavior and Organization, Vol. 162, pp. 251-272, doi: 10.1016/j.jebo.2018.12.021.
Zelazny, J. (2017), “Financialization and commodity market stability”, E-Finanse, Vol. 12, pp. 33-42, doi: 10.1515/fiqf-2016-0006.
Zhang, X., Yang, X., Li, J. and Hao, J. (2023), “Contemporaneous and noncontemporaneous idiosyncratic risk spillovers in commodity futures markets: a novel network topology approach”, Journal of Futures Markets, Vol. 43, pp. 705-733, doi: 10.1002/fut.22407.
Zhao, L. and Huchzermeier, A. (2017), “Integrated operational and financial hedging with capacity reshoring”, European Journal of Operational Research, Vol. 260, pp. 557-570, doi: 10.1016/j.ejor.2016.12.036.
Zhao, L.-T., Meng, Y., Zhang, Y.-J. and Li, Y.-T. (2019), “The optimal hedge strategy of crude oil spot and futures markets: evidence from a novel method”, International Journal of Finance and Economics, Vol. 24, pp. 186-203, doi: 10.1002/ijfe.1656.
Acknowledgements
Funding: The authors are grateful to the financial support provided by the National Natural Science Foundation of China (72273090).