COVID-19, the Russia-Ukraine war and the connectedness between the U.S. and Chinese agricultural futures markets

The paper examines how two major global crises—the COVID-19 pandemic and the Russia-Ukraine war—altered the connectedness between the U.S. and Chinese agricultural futures markets (soybean, corn, cotton). Motivated by increasing global integration of commodity markets and potential contagion during shocks, the study asks whether and how these crises changed correlation structures and lead–lag (causal) relationships, as well as volatility spillovers, across these markets. The U.S. is a key producer and exporter, while China is a major consumer and importer, making bilateral market linkages economically important for risk management, policy design, and price stability. The study is grounded in contagion theory, emphasizing information flows and financial linkages as channels for cross-border transmission of shocks, and focuses on short- and long-horizon dynamics relevant to investors and regulators.

Prior research has documented international linkages across commodities and between developed and emerging markets, especially during crises (e.g., Melvin and Sultan, 1990; Silvennoinen and Thorp, 2013; Mensi et al., 2019; Alquist et al., 2020). Studies explored specific cross-market ties such as U.S.–Canadian wheat (Booth et al., 1998) and London–Shanghai copper (Li and Zhang, 2008). Recent work shows COVID-19 increased interdependence between U.S.–Chinese stock and oil futures markets (Zhang and Mao, 2022; Zhang et al., 2022), and that COVID-19 and the Russia-Ukraine war had different effects on U.S.–Chinese stock market linkages (Zhang and Sun, 2023). However, evidence on international agricultural futures connectedness during these recent crises is limited. This paper addresses that gap by examining soybean, corn, and cotton futures linkages between the U.S. and China across time scales, adding to contagion and connectedness literature with multiscale analysis.

Data: Daily closing prices for U.S. soybean, corn, and cotton futures (NYMEX) and Chinese soybean, corn, and cotton futures (Dalian Futures Exchange) from 2018-03-26 to 2023-11-28. Returns are log differences: R_t = ln(P_t) − ln(P_{t−1}). Descriptive statistics indicate higher price levels and price standard deviations in China, but higher return volatilities in the U.S. Sample sizes for each series are 1,479 observations. Empirical strategy: 1) Long-run linkage: Johansen cointegration tests between U.S. and Chinese returns for each commodity; no cointegration is found. 2) Time–frequency comovement and causality: Wavelet coherence analysis (WCA) assesses time-varying coherence and phase (lead–lag) relations across investment horizons (periods in days). Arrow orientations indicate correlation sign and leader–follower dynamics at each time and frequency. 3) Volatility spillovers: Time-varying parameter VAR (TVP-VAR) connectedness approach (Cagli et al., 2023) to compute dynamic total connectedness index (TCI), directional spillovers (To/From), and net spillovers between U.S. and Chinese markets for each commodity. 4) Crisis-period causality and shock transmission (soybeans): VAR models are estimated for defined subperiods with optimal lags selected via LR/FPE/AIC/HQIC/SBIC. Periods: pre-COVID-19 (2018-03-26 to 2020-01-17), pure-COVID-19 (2020-01-18 to 2021-02-23), war period (2021-02-24 to 2023-11-28). VAR stability is confirmed via roots inside the unit circle. Impulse response functions (IRFs) and forecast error variance decompositions (FEVDs) quantify cross-market responses and variance contributions. Model details: - WCA interprets in-phase (positive) vs out-of-phase (negative) comovement and which market leads at specific horizons (1–512 days). - TVP-VAR connectedness yields average and time-varying spillovers; TCI summarizes overall connectedness. Average connectedness tables report contribution TO, FROM, NET, and TCI. - VAR(1) used for pure-COVID-19; VAR(4) for war period (based on selection criteria).

- No long-run cointegration between U.S. and Chinese returns for soybean, corn, or cotton. - WCA—Soybeans: Around early 2020 (COVID-19 onset) and spring 2022 (Russia-Ukraine war), short-horizon (≈1–12 days) coherence patterns and lead–lag directions change, alternating between in-phase and out-of-phase comovements, with shifts between U.S.-led and China-led dynamics. - WCA—Corn: COVID-19 shows little effect around early 2020. Around early 2022, arrows indicate a change from in-phase to out-of-phase comovement at 8–12 day horizons, with leadership largely unchanged. - WCA—Cotton: Short-horizon arrows predominantly upward/right-upward (U.S. leading) and rightward at longer horizons (>16 days), indicating positive correlations; both crises have little impact on short-horizon causality and retain positive long-horizon correlation. - TVP-VAR average volatility spillovers (Contribution TO others; FROM in parentheses): • Soybean: U.S. → China 3.1% (China → U.S. 0.8%); TCI ≈ 3.9%. • Corn: U.S. → China 4.6% (China → U.S. 1.9%); TCI ≈ 6.5%. • Cotton: U.S. → China 14.7% (China → U.S. 5.7%); TCI ≈ 20.5%. Across all three, U.S. markets are net transmitters of volatility to China. - TVP-VAR dynamics: TCI for soybeans fell near zero during March–April 2020 (pandemic onset), then rose into 2021; at the start of the Russia-Ukraine war, TCI increased sharply over a short time scale. Similar dynamics are observed for corn. Cotton exhibits minimal changes around both crises. - VAR/IRFs (Soybeans): • Pure-COVID-19: One s.d. shocks ≈ 0.014 (U.S.), 0.013 (China). Chinese returns respond notably to a U.S. shock (≈0.002), lasting ~2 days; U.S. response to a Chinese shock is close to zero. • War period: One s.d. shocks ≈ 0.015 (U.S.), 0.009 (China). Chinese returns respond more strongly and persistently to a U.S. shock (≈0.002, ~6 days), while the U.S. responds modestly to a Chinese shock (≈0.001, ~5 days). - FEVD (Soybeans): • COVID-19 period: Chinese variations contribute ≈0.05% to U.S. return error variance; U.S. variations contribute ≈2% to Chinese variance. • War period: Chinese → U.S. contribution rises to ≈0.3–0.5%; U.S. → China rises to ≈4.5%. - Overall: U.S. agricultural markets consistently transmit more volatility risk to China; the war’s spillover enhancement effect is stronger than the pandemic’s for soybeans and corn, with negligible crisis effects in cotton.

The findings indicate that exogenous shocks from COVID-19 and the Russia-Ukraine war reshape short-horizon comovements and causality between U.S. and Chinese agricultural futures, particularly for soybeans and corn, aligning with contagion theory via information flows and financial linkages. The pandemic initially muted cross-market volatility spillovers (TCI declined near zero), possibly due to lockdowns and temporary disruptions, before spillovers gradually recovered. The Russia-Ukraine war, given Ukraine’s role as a major agricultural exporter, sharply amplified short-horizon risk transmission, especially from U.S. to China. Cotton’s relative insensitivity suggests sector-specific demand characteristics (textile inputs being less urgent during crises) reduce contagion. Policy significance: Recognizing dominant U.S.→China volatility transmission can inform coordination on trade policies, monitoring, and targeted interventions to stabilize prices (e.g., managing quotas/tariffs, buffer stocks). Investor relevance: Cross-market hedging and diversification strategies should account for time-varying connectedness and stronger U.S.-origin spillovers, with heightened vigilance during geopolitical shocks. The multiscale perspective (WCA) helps align strategies with investment horizons.

The study shows that COVID-19 and the Russia-Ukraine war materially altered short-horizon correlation and lead–lag relations for U.S.–Chinese soybean and corn futures, with minimal effects for cotton. U.S. markets are net transmitters of volatility to China across all three commodities, and the war strengthened spillovers more than the pandemic. For soybeans, Chinese returns respond more strongly and persistently to U.S. shocks than vice versa, especially during the war. These results extend contagion theory by highlighting horizon-dependent transmission and inform policy and risk management during crises. Future research can apply the WCA and TVP-VAR framework to broader sets of commodities and additional international markets (e.g., G20, BRICS) to map global contagion patterns.