Next Article in Journal
Impacts of Emissions Trading Scheme Initiatives on Corporate Carbon Proactivity and Financial Performance
Next Article in Special Issue
When Institutional Plates Collide: The Dynamic Impact of Informal Institutions on Capital Market Development
Previous Article in Journal
Examining the Link between Technical Efficiency, Corporate Governance and Financial Performance of Firms: Evidence from Nigeria
Previous Article in Special Issue
The Worst Case GARCH-Copula CVaR Approach for Portfolio Optimisation: Evidence from Financial Markets
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Forecastability of Agricultural Commodity Futures Realised Volatility with Daily Infectious Disease-Related Uncertainty

1
Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa
2
Department of Rural Studies, Tripura University, Agartala 799022, Tripura, India
*
Author to whom correspondence should be addressed.
J. Risk Financial Manag. 2022, 15(11), 525; https://doi.org/10.3390/jrfm15110525
Submission received: 12 October 2022 / Revised: 3 November 2022 / Accepted: 8 November 2022 / Published: 10 November 2022
(This article belongs to the Special Issue Applied Financial Econometrics)

Abstract

:
Given the food supply chain disruption from COVID-19 lockdowns around the world, we examine the predictive power of daily infectious diseases-related uncertainty (EMVID) on commodity traded futures within the agricultural bracket, sometimes known as the softs, using the heterogeneous autoregressive realised variance (HAR-RV) model. Considering the short-, medium-, and long-run recursive out-of-sample estimation approach, we estimate daily realised volatility by using intraday data within the 5 min interval for 15 agricultural commodity futures. During the COVID-19 episode, our results indicated that EMVID plays an important role in predicting the future path of agricultural commodity traded futures in the short, medium, and long run, i.e., h = 1, 5, and 22, respectively. According to the MSE-F test, these results are statistically significant. These results contain important implications for investors, portfolio managers, and speculators when faced with investment risk management and strategic asset allocation during infectious disease-related uncertainty.

1. Introduction

The disruption of food supply chains from COVID-19 lockdowns around the world triggered a tremendous interest in understanding the “safe haven” attribute of agricultural commodity futures (Ji et al. 2020; Sifat et al. 2021; Rubbaniy et al. 2022; Zhang and Wang 2022), raising concerns about the attractiveness of these vehicles in commodity options trading, global supply chain risk management,1 strategic asset allocations, and regulators’ supervision of inflation risk during infectious disease-related uncertainty.
In 2015, the United Nations set 17Sustainable Development Goals (SDGs) that were aimed at improving the standard of living in the world by 2030 (SDSN 2021). Among these SDGs are those of no poverty (SDG1) and zero hunger (SDG2) by 2030. The COVID-19 outbreak imposed the greatest threat to these goals and adversely affected some of the developing progress in achieving them when governments imposed measures such as lockdowns2 to contain the spread of the virus (Khan et al. 2020). In addition to the lost lives, Béné (2020) emphasised that the main effect of COVID-19 was driven by mobility restrictions by governments, which led to a subsequent loss of income and reduction in purchasing power, especially for low-income individuals and households. The restricted movement between countries (see McBryde et al. 2020) triggered demand and supply shocks (Guerrieri et al. 2022). This threatened food security, the most crucial aspect of sustainable development and economic growth in different parts of the world (Arndt et al. 2020; Mardones et al. 2020; O’Hara and Toussaint 2021). Empirically, approximately 265 million people were affected by food insecurity in 2020, which is a 135 million increase from the COVID-19 outbreak (Food Security Information Network 2020).
The interest of our paper in commodity markets is driven by food security and their more dramatic price fluctuations compared with other financial markets (Hák et al. 2016). If we think of agricultural commodities, for instance, the production of goods is not uniform throughout the year (de Keizer et al. 2017). Crops, for example, grow in a certain season and are usually harvested a few times a year, and it can often be unpredictable up to a certain point whether the crops will turn out good or bad. The weather conditions have a big effect on these outcomes; however, we may have other unpredictable factors such as pesticides (Tudi et al. 2021). These kinds of fluctuations are a problem for commodity producers, investors, and portfolio managers. However, the COVID-19 outbreak led to high volatility as a result of the high unprecedented uncertainties in the financial market,3 especially in the commodity markets. Therefore, it is crucial for investors and portfolio managers to mitigate or offset such risk by finding “safe haven” commodity futures during times of infectious diseases.
In times of financial market uncertainties from global crises such as infectious diseases, especially the recent coronavirus pandemic, typically used portfolio risk management strategies are likely to default (Umar and Gubareva 2020; Harjoto et al. 2021). This may result in extreme market volatility because of high trades. More precisely, the disastrous COVID-19 pandemic prompted a high level of uncertainty in the commodity markets although the reaction of such markets differed across countries and traded commodity brackets. For instance, commodity-dependent countries rely heavily on exports and imports as low-and middle-income countries; as a result, they experienced a strong adverse reaction in their markets (Tröster 2020). On the other hand, Borgards et al. (2021) showed that the reaction of agricultural (soft) and metal commodities to the pandemic was minimal except for special treasures such as gold. In addition, Zhang and Hamori (2021) argued that the effects of COVID-19 on the financial markets are more significant compared to other historical shocks such as the 2008 financial crisis, droughts, and floods, although their short-, medium-, and long-run impact is uncertain.
In this context, the objective of our paper is to investigate, for the first time, the predictive ability of daily infectious disease-related uncertainty (EMVID) for agricultural future realised volatilities utilising the heterogeneous autoregressive realised variance (HAR-RV) model. The main attribute of the HAR-RV model is its ability to use volatilities from different time horizons to predict the realised volatility on returns. The model contains the heterogeneous market hypothesis, which states that market participants in their different categories react differently to information flow in the short, medium, and long run (Müller et al. 1997). For example, speculators and traders in the market are more concerned about short-term investments, while investors are more interested in long-term investments. Conventionally, the time-varying volatility is modelled, and the fit is assessed using various generalised autoregressive conditional heteroscedastic (GARCH) models, under which the conditional variance is a deterministic function of model parameters and past data. Alternatively, researchers have also considered stochastic volatility models, where the volatility is a latent variable that follows a stochastic process. These models rely on daily data, and not intraday data as used to obtain RV, which in turn is known to a be more accurate estimate of the latent process of volatility due to the richness of the underlying intraday data (McAleer and Medeiros 2008).
There are a number of studies on the nexus between commodity returns and infectious diseases, especially since the incidence of the COVID-19 pandemic (See Balcilar et al. 2022; Long and Guo 2022; Akyildirim et al. 2022; do Nascimento et al. 2022; Daglis et al. 2020; Umar et al. 2022; Cariappa et al. 2022; Chen et al. 2022; Shruthi and Ramani 2021; Gutierrez et al. 2022; Ayyildiz 2022). However, the current study makes key contributions to the existing literature. First, the focus of existing studies was mainly on the COVID-19 pandemic, while the current study focuses on infectious disease-related uncertainty (EMVID). Secondly, existing studies used daily data for commodity returns, while we use the realised volatility of intraday agricultural commodity futures. Thirdly, relative to existing studies, we analyse the out-of-sample power of EMVID for more (15) agricultural commodity futures (i.e., BO, CC, C, CT, KC, OJ, SB, SM, S, W, FC, LB, LC, LH, and O) (Table A1). The data coverage of uncertainty related to infectious diseases not only covers the COVID-19 episode, but also includes other infectious diseases such as Ebola, H1N1, H5N1, MERS, or SARS viruses and the recent monkeypox. We use the newspaper-based index by Baker et al. (2020) as a proxy for infectious disease-related uncertainty. The index is derived from the daily equity market volatility (EMV) hosted in the Chicago Board Options Exchange (CBOE) volatility index. This index is robust for a statistical model aimed at forecasting the realised volatility of agricultural commodity futures. Most importantly, the employed intraday data contains information that may result in more accurate and precise estimates and forecasts across different time horizons. Furthermore, this paper contributes to the literature on agricultural commodity futures in that it predicts its realised volatility computed from 5 min intervals utilising the modified heteroscedasticity autoregression model by Corsi (2009). In particular, the basic HAR-RV model is extended by adding the daily infectious disease-related uncertainty (EMVID) variable, and then examining its predictive power on the variables of interest (agricultural commodity futures). Furthermore, we employ recursive out-of-sample predictability of EMVID for the realised volatility of 15 agricultural commodity futures in the short, medium, and long run. In sum, our study is holistic and novel in terms of the wider coverage of infectious disease range, the focus on intraday realised volatility of large number agricultural commodities, the focus on the out-of-sample predictability of EMVID, and the uniqueness of the modified HAR-RV model used, allowing us to conduct short-, medium-, and long-run forecast analysis. To the best of our knowledge, we are not aware of any study that has examined the out-of-sample predictability of EMVID for the intraday volatility of agricultural commodities using the HAR-RV model. This analysis has important implications for portfolio managers in their portfolio diversification possibilities given uncertainties from infectious diseases.
The remaining part of our paper is structured as follows: Section 2 presents the literature review, while Section 3 describes the data and methodology. Section 4 presents the empirical results. Section 5 concludes the paper. Figure A1 presents the data plots for our variables of interest.

2. Literature Review

Several studies have been conducted on the infectious disease and financial markets, especially since the incidence of COVID-19 pandemic (see, for example, Salisu and Vo 2020; Salisu et al. 2020; Caggiano et al. 2020; Bouri et al. 2020b; Salisu and Sikiru 2020; Salisu and Adediran 2020; Salisu et al. 2020; Adediran et al. 2021; Liu et al. 2022). However, these studies focused mainly on stock returns. Some studies exist on commodity returns and infectious diseases. For example, using the nonparametric Granger causality-in-quantiles test, Balcilar et al. (2022), assessed the effect of COVID-19 (measured by the news-based sentiment index) on 13 major agricultural commodity prices and price volatility. They employed daily data over 73 months, i.e., from 1 January 2016 to 25 February 2022. Their findings suggest that in both the lower and upper quantile ranges, there is Granger causality from the pandemic to the average commodity prices. Furthermore, COVID-19 sentiment is causal to the price volatility of agricultural commodities in the quantiles above the first quarter.
Long and Guo (2022) analysed the effects of infectious disease equity market volatility and other factors on commodity returns. Results based on time-varying Granger causality test and time-varying parameter vector autoregression with a stochastic volatility model showed that the time-varying effects are significant with mostly positive responses. They also found out that, of the five pandemics (Bird Flu in 1998, SARS in 2003, Swine Flu in 2009, MERS and Ebola in 2014, and COVID-19 in 2019) studied, the recent COVID-19 produced the greatest impact on commodity returns. Furthermore, they showed that the returns of five commodity subcategories, namely, textiles, industry, metals, livestock, and food, were mostly negatively impacted during the sample period, thereby making these commodities not safe haven assets during pandemic risks.
Akyildirim et al. (2022) used panel data regressions and time-varying Granger causality tests to examine whether the spillovers between agricultural commodity returns and sentiments are influenced by economic and financial uncertainties, including the global COVID-19 pandemic. They found that the agricultural commodity returns and sentiments were significantly influenced by COVID-19-induced uncertainty around the first cycle of the pandemic in 2020. do Nascimento et al. (2022) used the Hurst exponent and multifractal detrended fluctuations analysis, and they found that, during the COVID-19 pandemic (from 1 January 2020 to 25 September 2020), sugar was the most efficient commodity, while pork was the least in the Brazilian agricultural commodity market. Daglis et al. (2020) analysed the impact of COVID-19 pandemic on oats and wheat returns using data from 22 January 2020 to 2 June 2020. Results from the standard VAR model indicated that these markets were affected by COVID-19. Furthermore, these results indicated the out-of-sample forecasting superiority of a model that explicitly incorporates COVID-19 pandemic over the baseline model.
Using data from 1 January 2020 and 30 April 2021, Umar et al. (2022) examined the dynamic return and volatility connectedness for three agricultural commodity indices (softs, grains, and livestock) and the coronavirus media coverage index (MCI). Results based on time-varying parameter vector autoregression showed that dynamic total return and volatility connectedness fluctuated over time, reaching a peak during both the first and the third waves of the COVID-19 pandemic. Cariappa et al. (2022) used time series data from 1 November 2019 to 10 August 2020 in conjunction with survey data to analyse the effect of COVID-19-induced lockdowns on agricultural commodity prices and consumer behaviour in India. Results from an interrupted time series analysis showed a significant rise in the prices of chickpea (4.8%), mung bean (5.2%), and tomato (78.2%), although these reverted immediately after the lockdown. Furthermore, the Kruskal–Wallis test results showed a significant change in consumer behaviour through panic purchases.
Chen et al. (2022) used data from 2019 to 2021 and the Black (1976) model to show how theCOVID-19 pandemic impacted the volatility of Chinese agricultural commodity options more strongly relative to non-agricultural commodities. Using causality in impulse response functions and variance test and daily data from January 2020, Shruthi and Ramani (2021) found that the risk transmission among agricultural commodities was zero. According to Gutierrez et al. (2022), results from a global vector auto regression (GVAR) model revealed that the fall in the oil price may have contributed to the stability of the world grain market in during COVID-19 pandemic and that export restrictions could significantly increase global prices. An asymmetric analysis by Ayyildiz (2022) using the nonlinear autoregressive distributed lag model and data from 11 March 2020 to 11 March 2021 showed that the effect of an increase in the COVID-19 global fear index on agricultural commodity prices was greater than the effect of a decrease.
According to the above, the majority of these studies focused on COVID-19 pandemic while the current study uses an infectious disease uncertainty index that is broad and covers different infectious diseases pandemics. Furthermore, all the studies except Daglis et al. (2020) conducted in-sample predictability analysis, while we conduct an out-of-sample analysis. It is, however, noted that, while Daglis et al. (2020) analysed the out-of-sample predictability of COVID-19 for oats and wheat returns, we analyse the out-of-sample predictability of infectious disease-related uncertainty for the realised volatility of 15 agricultural commodity futures. Hence, we innovate by focusing on volatilities in several agricultural commodities in the futures market.

3. Data and Methodology

3.1. Data

Data on the realised volatility (RV) of commodity futures were sourced directly from the University of Chicago Booth School of Business Risk Lab under the maintenance of Professor Dacheng Xiu. This series is publicly available at https://dachxiu.chicagobooth.edu/#risklab.com (accessed on 27 April 2022). The highest-frequency available trades were collected and cleaned using the prevalent national best bid and offer (NBBO) that is available every second. The RV estimation procedure was computed using the quasi-maximum likelihood estimation of volatility (QMLE) from moving average models MA(q), using nonzero returns of transaction prices sampled up to the earliest available frequency for days with at least 12 observations (see Xiu 2010). In choosing the best MA(q), we used the Akaike information criterion. We also employed the 5min RV estimates for our analysis.
The index on dairy infectious disease-related uncertainty (EMVID) is publicly accessible at http://policyuncertainty.com/infectious_EMV.html (accessed on 27 April 2022).This index was developed by Baker et al. (2020) using a newspaper-based infectious disease equity market volatility tracker. In this paper, we use the EMVID data from as early as 22 September 2008 to 27 April 2022 for BO, CC, C, CT, KC, OJ, SB, SM, S, and W RVs, and then from 27 July 2015 to 27 April 2022 for FC, LB, LC, LH, and O RVs (Table A1). EMVID is based on the following four textual analysis terms: E, economic, economy, financial; M, “stock market”, equity, equities, “standard and poor”; V: volatility, volatile, uncertain, uncertainty, risky; ID: H1N1, H5N1, MERS, SARS, Ebola pandemic, epidemic, virus, diseases, and coronavirus. In each of the E, M, V, and ID terms, a daily count of at least one term over 3000 US newspaper articles was computed into the EMVID index. On the same day, Baker et al. (2020) multiplicatively rescaled the final series to equal the level of the VIX through the overall EMV index; then, the EMVID index was scaled to total the EMV articles. Our data range varied from the earliest data available to the latest date from our estimation. More interestingly, our data period covers the COVID-19 virus and other economic uncertainties such as the global financial crisis and, more recently, the Russia–Ukraine crisis. Given daily infectious disease-related uncertainty, the EMVID index is the only proxy for uncertainty related to various infectious diseases.

3.2. Methodology: Heterogeneous Autoregressive Realised Variance (HAR-RV) Model

To realise the main objective of our paper, we conducted the out-of-sample predictability analysis using the Corsi (2009) HAR-RV model. The key feature of our model is its ability to reproduce the important properties contained in financial data in their respective time intervals while remaining simple (Wang et al. 2019; Gkillas et al. 2020). These properties include fat tails, long memory, multi-scaling behaviour, and self-similarity. The basic HAR-RV model is
R V t + h = β 0 + β d R V t + β w R V w . t + β m R V m . t + ε t + h ,
where realised volatility (RV) h days ahead is represented by the h index (in our paper, h = 1, 5, and 22); R V w . t represents the average RV from day t − 6 to t − 1, whereas R V m . t depicts the mean RV from day t − 22 to day t − 6. We then add the EMVID index to the benchmark HAR-RV model to capture the interest of our paper. β 0 is a constant, ceteris paribus. β d , w   a n d   m are our respective coefficients for the short-, medium-, and long-run RV, while ε t + h is our error term. The extended HAR-RV model ( θ is the coefficient for daily infectious disease-related uncertainty) is
R V t + h = β 0 + β d R V t + β w R V w . t + β m R V m . t + θ E M V I D t + ε t + h .

4. Empirical Results

In this paper, we focus on the out-of-sample predictability of the realised volatility (RV) of commodity traded futures, “the softs”; that is, we access the role that daily infectious disease-related uncertainty (EMVID) plays in predicting the future path of our variables of interest. Campbell (2008) and Bouri et al. (2020a) argued that the best test for any predictive model relies on its out-of-sample performance in terms of any econometric and predictability. We employ an out-of-sample recursive approach from the earliest data available to the latest data for our estimation. The data plots on the variables under investigation in Figure A1 move around the mean with a sharp positive shock that quickly goes back to the mean in the first quarter of the COVID-19 pandemic, especially for our independent variable. Our out-of-sample multiple structural breakpoints tests were determined using the HAR-RV model under the Bai and Perron (2003) test of 1 to M globally determined breaks and UDMax and WDMax statistics. Table 1 presents the structural breaks.
As tested by the multiple structural breakpoints test, Table 1 depicts that most agricultural commodity futures experienced multiple structural breaks in 2010. More precisely, corn (C), cotton #2 (CT), and soybean (S) futures experienced a structural breakpoint in October 2010, followed by soybean oil (BO) and live cattle (LC) futures in November 2010. The orange juice (OJ) and sugar #11 (SB) futures had a structural breakpoint in December 2010. In September 2010, the coffee “C” (KC) experienced a breakpoint. Furthermore, cocoa (CC) and soybean meal (SM) futures had a structural breakpoint in March 2011, and wheat futures CBOT (W) experienced a breakpoint in October 2011. Lastly, the feeder cattle (FC), lumber (LB) and oats (O) futures had a structural breakpoint in August 2016, and the lean hogs (LH) futures experienced a breakpoint in October 2016. The important basis of these multiple structural breakpoints involves factors such as food price peaking, reduction in grain stock, low interest rates, and the depreciation of the United States (US) dollar (Headey 2011). Export restrictions, droughts, demand surges, trade shocks, and climate change are among other factors contributing to the global food crisis (see Falkendal et al. 2021; Lieber et al. 2022).
Next, we compute the root-mean-squared forecast errors (RMSFEs) for the benchmark and extended h = 1, 5, and 22 HAR-RV models using the above multiple structural breakpoints models. Since our primary aim is to forecast, lower RMSFEs in our recursive out-of-sample estimated from the earliest experienced breakpoint in all the variables of interest would represent a better-performing model. For forecast accuracy, we employ the McCracken (2007) MSE-F test. The out-of-sample forecast gains (FG) were calculated using the following formula:
F G = ( R M S F E 0 R M S F E 1 1 ) × 100 ,
where RMSFE0 denotes the RMSFEs for the benchmark HAR-RV model, while the RMSFEs for the extended HAR-RV model are presented by RMSFE1. Positive or negative FGs indicate the gains or losses in percentage (Equation (3)).
According to our out-of-sample results in Table 2, the highest forecast loss of 0.28% was for the lumber futures (LB), followed by 0.26% forecast loss for soybean oil futures (BO) in the short run (h = 1), and then 0.25% in the medium run (h = 5) in the BO. This implies that taking the information context of the daily infectious disease-related uncertainty (EMVID) into consideration using the forecast accuracy of the RMSFE metrics within our period of interest, an econometrician can obtain the highest forecast loss of 0.28% for LB (h = 1), followed by 0.26% and then 0.25% for BO h = 1 and h = 5, respectively. Our results also indicate that the coffee “C” (h = 22) and oat futures (O) (h = 5) remained constant, i.e., there was no forecast gain or loss. However, the lowest forecast loss of 0.01% was in the oat futures h = 1 model, followed by 0.02 for wheat futures CBOT (W) in the h = 22 model. This suggests that, considering the information context of uncertainty associated with infectious diseases based on the forecast accuracy of the RMSFE metrics, an econometrician would not be able to obtain any forecast gain or loss for KC (h = 22) and O (h = 5), but could at least obtain a minimal forecast loss of 0.01% for O (h = 1), followed by 0.02% for W (h = 22). Considering the whole sample period, these negative FGs also imply that EMVID adds no value in forecasting the realised volatility of our commodity futures. Therefore, the MSE-F test cannot be significant it is a one-sided test associated with whether the unrestricted model does better than the restricted one.
Across all economic agents, the interest in searching for “safe haven” vehicles given infectious disease-related uncertainty was triggered by the COVID-19 outbreak; therefore, it is crucial to assess the impact of EMVID within the COVID-19 period. As the primary purpose of this paper, we conducted a recursive out-of-sample estimation from January 2020 to the earliest period of our estimation and computed the in-sample period including the same number of observations. That is, we performed in- and out-of-sample observations. This period incorporates all phases of COVID-19. Within the COVID-19 episode, our results in Table 3 depict that the cocoa futures (CC) had the highest FG of 265.12% in the h = 22 model, followed by 119.38% for oats futures in the h = 22 model, and then 91.40% for sugar #11 futures (h = 1). This implies that, by incorporating the information context of infectious disease-related uncertainty such as COVID-19 using the forecast accuracy of the RMSFE metrics, an econometrician could acquire the highest FG of 265.12% for CC (h = 22), followed by 119.37% for O (h = 22), and then 91.40 for SB (h = 1). Furthermore, within the same episode, the lowest forecast gains of 0.70%, 1.49%, and 1.68%were evident in the SM (h = 5), SB (h = 22), and SM (h = 1), respectively. This means that, considering COVID-19-related uncertainty and the forecast accuracy RMSFE metrics, an econometrician could obtain the lowest FGs of 0.70% in SM (h = 5), followed by 1.49% for SB (h = 22), and then 1.68% for SM (h = 1). According to the MSE-F critical values,4 these results were statistically significant at a 1% level of significance except for BO in the h = 1 and h = 5 models. Most importantly, the results of our out-of-sample in the COVID-19 episode indicate the extent to which trade openness can be affected by a national shutdown given infectious diseases. Specifically, the supply shock triggered food insecurity; as a result, there was a high willingness to hedge against such risks.

5. Conclusions

Given food insecurity problems as a result of the COVID-19 lockdowns around the world, we investigated the forecasting ability of daily infectious disease-related uncertainty (EMVID) with respect to the realised volatility of agricultural commodity traded futures. We employed the heterogeneous autoregressive realised variance (HAR-RV) model by Corsi (2009) on 15 commodity traded futures. Considering our recursive out-of-sample estimation approach in the short, medium, and long run within the COVID-19 episode, it is evident that cocoa futures (CC) had the highest FG of 265.12% in the long run (h = 22), followed by oat futures (O) with 119.38% FG in h = 22, and then 91.40% FG for sugar #11 (SB) in the short run (h = 1). This implies that, considering the information context of the forecasting accuracy for RMSFE metrics within the COVID-19 period, an econometrician could obtain the highest FG of 265.12% in CC h = 22, followed by 119.38% for O h = 22, and then 91.40% for SB h = 1. An econometrician could also obtain the lowest FG of 0.70%, followed by 1.49% and 1.68% in SM h = 5, SB h = 22, and SM h = 1, respectively.
Our results within the COVID-19 episode suggest that EMVID plays an important role in predicting the future path of agricultural commodity futures. These findings have important implications for portfolio managers and investors in their search for safe investment or diversification options in the financial market. These results are robust as suggested by McCracken’s (2007) MSE-F test. The COVID-19 pandemic is the worst crisis the world had seen; therefore, there are limited related studies and measures or indices for COVID-19. Furthermore, the pandemic already aggravated existing food insecurity problems and other global challenges; hence, we cannot blame the volatility of this asset class under review solely on the pandemic. In the future, we expect to extend our study to other brackets of agricultural commodities such as those in the metal bracket.

Author Contributions

Conceptualization, S.S., G.C.A., R.G. and S.G.; methodology, S.S.; validation, S.S., G.C.A., R.G. and S.G.; formal analysis S.S., G.C.A., R.G. and S.G.; writing—original draft preparation, S.S., G.C.A., R.G. and S.G.; writing—review and editing, S.S., G.C.A., R.G. and S.G. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Data are available under request from the authors, but the raw data is publicly available as stated in the data segment.

Acknowledgments

We thank the anonymous referees; any remaining errors are solely our responsibility.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Table A1. Selected variables, acronyms, and sample coverage.
Table A1. Selected variables, acronyms, and sample coverage.
SymbolFuture IndexSample Period
1. BOSoybean oil futures22 September 2008–27 April 2022
2. CCCocoa futures22 September 2008–27 April 2022
3. CCorn futures22 September 2008–27 April 2022
4. CTCotton no.2 futures22 September 2008–27 April 2022
5. FCFeeder cattle futures27 September 2015–27 April 2022
6. KCCoffee c futures22 September 2008–27 April 2022
7. LBLumber futures27 July 2015–27 April 2022
8. LCLive cattle futures27 July 2015–27 April 2022
9. LHLean hogs futures27 July 2015–27 April 2022
10. OJOrange juice futures22 September 2008–27 April 2022
11. OOat futures27 July 2015–27 April 2022
12. SBSugar #11 futures22 September 2008–27 April 2022
13. SM Soybean meal futures 22 September 2008–27 April 2022
14. SSoybean futures22 September 2008–27 April 2022
15. WWheat futures CBOT22 September 2008–27 April 2022
Figure A1. Data plots. Notes: The realised volatility of the agricultural commodity futures is represented by RV. The newspaper-based uncertainty index related to infectious disease is represented by EMVID.
Figure A1. Data plots. Notes: The realised volatility of the agricultural commodity futures is represented by RV. The newspaper-based uncertainty index related to infectious disease is represented by EMVID.
Jrfm 15 00525 g0a1aJrfm 15 00525 g0a1b

Notes

1
The profitability of businesses heavily depends on risk management strategies to hedge futures cash flow uncertainty.
2
Lockdowns reduced the movement of goods and services and even brought some to zero, i.e., movements of imports and exports.
3
Liao et al. (2018) noted the following three channels through which the fluctuation in the financial market can impact commodity prices: macro-economy reflection channel, financial market information transmission channel, and market sentiment contagion channel.
4
MSE-F critical values: 3.584, 1.548, and 0.751.

References

  1. Adediran, Idris A., Olalekan D. Yinusa, and Kanwal Hammad Lakhani. 2021. Where lies the silver lining when uncertainty hang dark clouds over the global financial markets? Research Policy 70: 101932. [Google Scholar] [CrossRef]
  2. Akyildirim, Erdinc, Oguzhan Cepni, Linh Pham, and Gazi Salah Uddin. 2022. How connected is the agricultural commodity market to the news-based investor sentiment? Energy Economics 113: 1–22. [Google Scholar] [CrossRef]
  3. Arndt, Channing, Rob Davies, Sherwin Gabriel, Laurence Harris, Konstantin Makrelov, Sherman Robinson, Stephanie Levy, Witness Simbanegavi, Dirk van Seventer, and Lillian Anderson. 2020. COVID-19 lockdowns, income distribution, and food security: An analysis for South Africa. Global Food Security 26: 100410. [Google Scholar] [CrossRef]
  4. Ayyildiz, Merve. 2022. Asymmetrical relationship between COVID-19 global fear index and agricultural commodity prices. Emirates Journal of Food and Agriculture 34: 239–47. [Google Scholar] [CrossRef]
  5. Bai, Jushan, and Pierre Perron. 2003. Computation and analysis of multiple structural change models. Journal of Applied Econometrics 18: 1–22. [Google Scholar] [CrossRef] [Green Version]
  6. Baker, Scott R., Nicholas Bloom, Steven J. Davis, and Stephen J. Terry. 2020. COVID-Induced Economic Uncertainty. Cambridge: National Bureau of Economic Research. [Google Scholar]
  7. Balcilar, Mehmet, Kamil Sertoglu, and Busra Agan. 2022. The COVID-19 Effects on Agricultural Commodity Markets. Agrekon 61: 239–65. [Google Scholar] [CrossRef]
  8. Béné, Christophe. 2020. Resilience of local food systems and links to food security—A review of some important concepts in the context of COVID-19 and other shocks. Food Security 12: 805–22. [Google Scholar] [CrossRef]
  9. Black, Fischer. 1976. The pricing of commodity contracts. Journal of Financial Economics 3: 167–79. [Google Scholar] [CrossRef]
  10. Borgards, Oliver, Robert L. Czudaj, and Thi Hong Van Hoang. 2021. Price overreactions in the commodity futures market: An intraday analysis of the Covid-19 pandemic impact. Resources Policy 71: 101966. [Google Scholar] [CrossRef]
  11. Bouri, Elie, Riza Demirer, Rangan Gupta, and Christian Pierdzioch. 2020a. Infectious diseases, market uncertainty and oil market volatility. Energies 13: 4090. [Google Scholar] [CrossRef]
  12. Bouri, Elie, Syed Jawad Hussain Shahzad, David Roubaud, Ladislav Kristoufek, and Brian Lucey. 2020b. Bitcoin, gold, and commodities as safe havens for stocks: New insight through wavelet analysis. The Quarterly Review of Economics and Finance 77: 156–64. [Google Scholar] [CrossRef]
  13. Caggiano, Giovanni, Efrem Castelnuovo, and Richard Kima. 2020. The global effects of Covid-19-induced uncertainty. Economics Letters 194: 109392. [Google Scholar] [CrossRef] [PubMed]
  14. Campbell, John Y. 2008. Estimating the equity premium. Canadian Journal of Economics/Revue Canadienne D’économique 41: 1–21. [Google Scholar] [CrossRef] [Green Version]
  15. Cariappa, A. G. Adeeth, Kamlesh Kumar Acharya, Chaitanya Ashok Adhav, Ramadas Sendhil, and P. Ramasundaram. 2022. COVID-19 induced lockdown effects on agricultural commodity prices and consumer behaviour in India—Implications for food loss and waste management. Socio-Economic Planning Sciences 82: 101160. [Google Scholar] [CrossRef]
  16. Chen, Jilong, Liao Xu, and Hao Xu. 2022. The impact of COVID-19 on commodity options market: Evidence from China. Economic Modelling 116: 105998. [Google Scholar] [CrossRef] [PubMed]
  17. Corsi, Fulvio. 2009. A simple approximate long-memory model of realized volatility. Journal of Financial Econometrics 7: 174–96. [Google Scholar] [CrossRef]
  18. Daglis, Theodoros, Konstantinos N. Konstantakis, and Panayotis G. Michaelides. 2020. The impact of COVID-19 on agriculture: Evidence from oats and wheat markets. Studies in Agricultural Economics 122: 132–39. [Google Scholar]
  19. de Keizer, Marlies, Renzo Akkerman, Martin Grunow, Jacqueline M. Bloemhof, Rene Haijema, and Jack G. A. J. van der Vorst. 2017. Logistics network design for perishable products with heterogeneous quality decay. European Journal of Operational Research 262: 535–49. [Google Scholar] [CrossRef]
  20. do Nascimento, Kerolly Kedma Felix, Fã Bio Sandro Dos Santos, Kellyma Kellyashin Felix Do Nascimento, Kenikywaynne Kerowaynne Felix Do Nascimento, Silvio Fernando Alves Xavier Jãšnior, Jader Silva Jale, and Tiago A. E. Ferreira. 2022. COVID-19 Influence over Brazilian Agricultural Commodities and Dollar–Real Exchange. Fractals 30: 2250100. [Google Scholar] [CrossRef]
  21. Falkendal, Theresa, Christian Otto, Jacob Schewe, Jonas Jägermeyr, Megan Konar, Matti Kummu, Ben Watkins, and Michael J. Puma. 2021. Grain export restrictions during COVID-19 risk food insecurity in many low-and middle-income countries. Nature Food 2: 11–14. [Google Scholar] [CrossRef]
  22. Food Security Information Network. 2020. Global Report on Food Crises: Joint Analysis for Better Decisions. Available online: https://docs.wfp.org/api/documents/WFP0000114546/download/?_ga=2.18682361.1730104652.1592430424-175228185.15924304242020 (accessed on 2 October 2022).
  23. Gkillas, Konstantinos, Rangan Gupta, and Christian Pierdzioch. 2020. Forecasting realized gold volatility: Is there a role of geopolitical risks? Finance Research Letters 35: 101280. [Google Scholar] [CrossRef]
  24. Guerrieri, Veronica, Guido Lorenzoni, Ludwig Straub, and Iván Werning. 2022. Macroeconomic implications of COVID-19: Can negative supply shocks cause demand shortages? American Economic Review 112: 1437–74. [Google Scholar] [CrossRef]
  25. Gutierrez, Luciano, Guillaume Pierre, and Maria Sabbagh. 2022. Agricultural Grain Markets in the COVID-19 Crisis, Insights from a GVAR Model. Sustainability 14: 9855. [Google Scholar] [CrossRef]
  26. Hák, Tomáš, Svatava Janoušková, and Bedřich Moldan. 2016. Sustainable Development Goals: A need for relevant indicators. Ecological Indicators 60: 565–73. [Google Scholar]
  27. Harjoto, Maretno Agus, Fabrizio Rossi, and John K. Paglia. 2021. COVID-19: Stock market reactions to the shock and the stimulus. Applied Economics Letters 28: 795–801. [Google Scholar] [CrossRef]
  28. Headey, Derek. 2011. Rethinking the global food crisis: The role of trade shocks. Food Policy 36: 136–46. [Google Scholar] [CrossRef]
  29. Ji, Qiang, Dayong Zhang, and Yuqian Zhao. 2020. Searching for safe-haven assets during the COVID-19 pandemic. International Review of Financial Analysis 71: 101526. [Google Scholar] [CrossRef]
  30. Khan, Karamat, Huawei Zhao, Han Zhang, Huilin Yang, Muhammad Haroon Shah, and Atif Jahanger. 2020. The impact of COVID-19 pandemic on stock markets: An empirical analysis of world major stock indices. The Journal of Asian Finance, Economics, and Business 7: 463–74. [Google Scholar] [CrossRef]
  31. Liao, Jia, Qi Qian, and Xiangyun Xu. 2018. Whether the fluctuation of China’s financial markets have impact on global commodity prices? Physica A: Statistical Mechanics and its Applications 503: 1030–40. [Google Scholar] [CrossRef]
  32. Lieber, Mark, Peter Chin-Hong, Knox Kelly, Madhavi Dandu, and Sheri D. Weiser. 2022. A systematic review and meta-analysis assessing the impact of droughts, flooding, and climate variability on malnutrition. Global Public Health 17: 68–82. [Google Scholar] [CrossRef]
  33. Liu, Wenwen, Yiming Gui, and Gaoxiu Qiao. 2022. Dynamics lead-lag relationship of jumps among Chinese stock index and futures market during the Covid-19 epidemic. Research in International Business and Finance 61: 101669. [Google Scholar] [CrossRef] [PubMed]
  34. Long, Shaobo, and Jiaqi Guo. 2022. Infectious disease equity market volatility, geopolitical risk, speculation, and commodity returns: Comparative analysis of five epidemic outbreaks. Research in International Business and Finance 62: 101689. [Google Scholar] [CrossRef] [PubMed]
  35. Mardones, Fernando O., Karl M. Rich, Lisa A. Boden, Andrea I. Moreno-Switt, Marisa L. Caipo, Natalia Zimin-Veselkoff, Abdulaziz M. Alateeqi, and Isabelle Baltenweck. 2020. The COVID-19 pandemic and global food security. Frontiers in Veterinary Science 7: 578508. [Google Scholar] [CrossRef]
  36. McAleer, Michael, and Marcelo C. Medeiros. 2008. Realized volatility: A review. Econometric Reviews 27: 10–45. [Google Scholar] [CrossRef]
  37. McBryde, Emma S., Michael T. Meehan, Oyelola A. Adegboye, Adeshina I. Adekunle, Jamie M. Caldwell, Anton Pak, Diana P. Rojas, Bridget M. Williams, and James M. Trauer. 2020. Role of modelling in COVID-19 policy development. Paediatric Respiratory Reviews 35: 57–60. [Google Scholar] [CrossRef] [PubMed]
  38. McCracken, Michael W. 2007. Asymptotics for out of sample tests of Granger causality. Journal of Econometrics 140: 719–52. [Google Scholar] [CrossRef]
  39. Müller, Ulrich A., Michel M. Dacorogna, Rakhal D. Davé, Richard B. Olsen, Olivier V. Pictet, and Jacob E. Von Weizsäcker. 1997. Volatilities of different time resolutions—Analyzing the dynamics of market components. Journal of Empirical Finance 4: 213–39. [Google Scholar] [CrossRef]
  40. O’Hara, Sabine, and Etienne C. Toussaint. 2021. Food access in crisis: Food security and COVID-19. Ecological Economics 180: 106859. [Google Scholar] [CrossRef]
  41. Rubbaniy, Ghulame, Ali Awais Khalid, Konstantinos Syriopoulos, and Aristeidis Samitas. 2022. Safe-haven properties of soft commodities during times of COVID-19. Journal of Commodity Markets 27: 100223. [Google Scholar] [CrossRef]
  42. Salisu, Afees A., and Abdulsalam Abidemi Sikiru. 2020. Pandemics and the Asia-Pacific Islamic stocks. Asian Economics Letters 1: 17413. [Google Scholar] [CrossRef]
  43. Salisu, Afees A., and Xuan Vinh Vo. 2020. Predicting stock returns in the presence of COVID-19 pandemic: The role of health news. International Review of Financial Analysis 71: 101546. [Google Scholar] [CrossRef]
  44. Salisu, Afees A., and Idris Adediran. 2020. Uncertainty due to infectious diseases and energy market volatility. Energy Research 1: 14185. [Google Scholar] [CrossRef]
  45. Salisu, Afees A., Lateef Akanni, and Ibrahim Raheem. 2020. The COVID-19 global fear index and the predictability of commodity price returns. Journal of Behavioral and Experimental Finance 27: 100383. [Google Scholar] [CrossRef] [PubMed]
  46. Shruthi, M. S., and D. Ramani. 2021. Statistical analysis of impact of COVID 19 on India commodity markets. Materials Today: Proceedings 37: 2306–231. [Google Scholar] [CrossRef]
  47. Sifat, Imtiaz, Abdul Ghafoor, and Abdollah Ah Mand. 2021. The COVID-19 pandemic and speculation in energy, precious metals, and agricultural futures. Journal of Behavioral and Experimental Finance 30: 100498. [Google Scholar] [CrossRef]
  48. Sustainable Development Solutions Network (SDSN). 2021. United States Sustainable Development Report 2021. United States of America. Available online: https://s3.amazonaws.com/sustainabledevelopment.report/2021/United+States+Sustainable+Development+Report+2021.pdf (accessed on 25 July 2022).
  49. Tröster, Bernhard. 2020. Commodity-Dependent Countries in the COVID-19 Crisis. Briefing Paper. Vienna: Austrian Foundation for Development Research (ÖFSE). [Google Scholar]
  50. Tudi, Muyesaier, Huada Daniel Ruan, Li Wang, Jia Lyu, Ross Sadler, Des Connell, Cordia Chu, and Dung Tri Phung. 2021. Agriculture development, pesticide application and its impact on the environment. International Journal of Environmental Research and Public Health 18: 1112. [Google Scholar] [CrossRef]
  51. Umar, Zaghum, and Mariya Gubareva. 2020. A time–frequency analysis of the impact of the COVID-19 induced panic on the volatility of currency and cryptocurrency markets. Journal of Behavioral and Experimental Finance 28: 100404. [Google Scholar] [CrossRef]
  52. Umar, Zaghum, Francisco Jareño, and Ana Escribano. 2022. Dynamic return and volatility connectedness for dominant agricultural commodity markets during the COVID-19 pandemic era. Applied Economics 54: 1030–54. [Google Scholar] [CrossRef]
  53. Wang, Xinya, Huifang Liu, Shupei Huang, and Brian Lucey. 2019. Identifying the multiscale financial contagion in precious metal markets. International Review of Financial Analysis 63: 209–19. [Google Scholar] [CrossRef]
  54. Xiu, Dacheng. 2010. Quasi-maximum likelihood estimation of volatility with high frequency data. Journal of Econometrics 159: 235–50. [Google Scholar] [CrossRef]
  55. Zhang, Wenting, and Shigeyuki Hamori. 2021. Crude oil market and stock markets during the COVID-19 pandemic: Evidence from the US, Japan, and Germany. International Review of Financial Analysis 74: 101702. [Google Scholar] [CrossRef]
  56. Zhang, Yongmin, and Ruizhi Wang. 2022. COVID-19 impact on commodity futures volatilities. Finance Research Letters 47: 102624. [Google Scholar] [CrossRef] [PubMed]
Table 1. Structural breakpoints.
Table 1. Structural breakpoints.
DateSymbolNames
September 2010KCCoffee “c” futures
October 2010C, CT, and SCorn futures, cotton #2 futures, and soybean futures
November 2010BO and LCSoybean oil futures and live cattle futures
December 2010OJ and SBOrange juice futures and sugar #11 futures
March 2011CC and SMCocoa futures and soybean meal futures
October 2011WWheat futures cbot
August 2016FC, LB, and OFeeder cattle futures, lumber futures, and oats futures
October 2016LHLean hogs futures
Table 2. Full out-of-sample forecasting gains.
Table 2. Full out-of-sample forecasting gains.
hRMSE0RMSEE1FGsRMSE0RMSEE1FGs
Panel 1: BO: 11/18/2010Panel 2: CC: 3/08/2011
10.04150.0416−0.26430.04970.0497−0.0282
50.01070.0107−0.25280.01310.0131−0.0229
220.00270.0027−0.07410.00330.0033−0.0302
Panel 3: C: 10/29/2010Panel 4: CT: 10/12/2010
10.07810.0781−0.07170.06320.0633−0.1201
50.02020.0202−0.05940.01650.0165−0.0666
220.00490.0049−0.06160.00410.0041−0.0978
Panel 5: FC: 8/18/2016Panel 6: KC: 9/14/2010
10.05030.0503−0.08750.05780.0578−0.1124
50.01290.0129−0.03100.01520.0152−0.0721
220.00180.0018−0.11030.00380.00380.0000
Panel 7: LB: 8/19/2016Panel 8: LC: 11/04/2010
10.17570.1762−0.27980.05310.0532−0.1937
50.04550.0455−0.13830.01310.0131−0.0915
220.01130.0114−0.16740.00350.0035−0.1442
Panel 9: LH: 10/11/2016Panel 10: OJ: 12/29/2010
10.07400.0740−0.08110.12390.1240−0.0468
50.01840.0184−0.07600.03240.0324−0.0309
220.00490.0049−0.12330.00780.0078−0.0385
Panel 11: O: 8/17/2016Panel 12: SB: 12/30/2010
10.13940.1394−0.00650.05700.0570−0.0526
50.03660.0366−0.00270.01480.0148−0.0271
220.00870.0087−0.03450.00370.0037−0.0540
Panel 13: SM: 3/21/2011Panel 11: S: 10/21/2010
10.05360.0536−0.02800.04610.0461−0.0390
50.01390.0139−0.03590.01200.0120−0.0334
220.00340.0034−0.02910.00290.0029−0.0344
Panel 15: W: 10/03/2011
10.06830.0684−0.0702
50.01830.0184−0.0436
220.00460.0046−0.0219
Note: F G = ( R M S F E 0 R M S F E 1 1 ) × 100 was the formula used to calculate the forecasting gains (FG), where R M S F E 0 stands for the root-mean-squared forecast errors ( R M S F E S ) for the benchmark model, and R M S F E 1 represents the R M S F E S for the extended HAR-RV model. R V t + h = β 0 + β d R V t + β w R V w , t + β m R V m , t + ε t + h is the equation for the benchmark HAR-RV model, and R V t + h = β 0 + β d R V t + β w R V w , t + β m R V m , t + θ E M V I D t + ε t + h is the equation for the extended HAR-RV model. RV depicts the daily realised volatility for agricultural commodity futures, whilethe daily infectious disease-related uncertainty is shown by EMVID.
Table 3. COVID-19 episode out-of-sample forecasting gains.
Table 3. COVID-19 episode out-of-sample forecasting gains.
hRMSE0RMSEE1FGsRMSE0RMSEE1FGs
Panel 1: BO: 01/02/2019Panel 2: CC: 01/02/2019
10.05930.0668−11.24640.06830.044154.9028 ***
50.01510.0164−8.04900.01670.011348.3345 ***
220.00390.00368.8284 ***0.01070.0029265.1210 ***
Panel 3: C: 01/02/2019Panel 4: CT: 01/02/2019
10.12250.074364.7963 ***0.11410.067369.4641 ***
50.03360.019572.1085 ***0.01950.017213.8156 ***
220.00580.004918.4884 ***0.00460.00444.2970 ***
Panel 5: FC: 01/02/2019Panel 6: KC: 01/02/2019
10.06190.052318.4753 ***0.07870.070311.9315 ***
50.02030.016919.9965 ***0.02450.018234.4523 ***
220.00260.002311.3804 ***0.00500.00476.8548 ***
Panel 7: LB: 01/02/2019Panel 8: LC: 01/02/2019
10.28230.249213.3014 ***0.10380.073740.7548 ***
50.08610.064134.3329 ***0.02390.017933.5645 ***
220.01650.01612.4484 ***0.00500.00482.9724 ***
Panel 9: LH: 01/02/2019Panel 10: OJ: 01/02/2019
10.09770.087112.2664 ***0.15520.131018.5122 ***
50.03020.021143.3042 ***0.04080.033521.8513 ***
220.00630.00596.7586 ***0.01240.007956.6002 ***
Panel 11: O:01/02/2019Panel 12: SB:01/02/2019
10.23430.144961.6987 ***0.09900.051791.4020 ***
50.03780.03702.2160 ***0.01990.013152.1473 ***
220.01970.0090119.3689 ***0.00340.00341.4784 ***
Panel 13: SM: 01/02/2019Panel 14: S: 01/02/2019
10.05170.05081.6788 ***0.06230.045836.0493 ***
50.01330.01330.7018 ***0.01520.011827.9527 ***
220.00600.003572.6407 ***0.00300.00300.0000 ***
Panel 15: W: 01/02/2019
10.16120.093572.4723 ***
50.04290.025766.8597 ***
220.00840.006332.7129 ***
Note: F G = ( R M S F E 0 R M S F E 1 1 ) × 100 was the formula used to calculate the forecasting gains (FG), where R M S F E 0 stands for the root-mean-squared forecast errors ( R M S F E S ) for the benchmark model, and R M S F E 1 represents the R M S F E S for the extended HAR-RV model. R V t + h = β 0 + β d R V t + β w R V w , t + β m R V m , t + ε t + h is the equation for the benchmark HAR-RV model, and R V t + h = β 0 + β d R V t + β w R V w , t + β m R V m , t + θ E M V I D t + ε t + h is the equation for the extended HAR-RV model. RV depicts the daily realised volatility for agricultural commodity futures, while the daily infectious disease-related uncertainty is shown by EMVID. The MSE-F test denotes the level of significance at the 1% level, as represented by ***.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Shiba, S.; Aye, G.C.; Gupta, R.; Goswami, S. Forecastability of Agricultural Commodity Futures Realised Volatility with Daily Infectious Disease-Related Uncertainty. J. Risk Financial Manag. 2022, 15, 525. https://doi.org/10.3390/jrfm15110525

AMA Style

Shiba S, Aye GC, Gupta R, Goswami S. Forecastability of Agricultural Commodity Futures Realised Volatility with Daily Infectious Disease-Related Uncertainty. Journal of Risk and Financial Management. 2022; 15(11):525. https://doi.org/10.3390/jrfm15110525

Chicago/Turabian Style

Shiba, Sisa, Goodness C. Aye, Rangan Gupta, and Samrat Goswami. 2022. "Forecastability of Agricultural Commodity Futures Realised Volatility with Daily Infectious Disease-Related Uncertainty" Journal of Risk and Financial Management 15, no. 11: 525. https://doi.org/10.3390/jrfm15110525

Article Metrics

Back to TopTop