Trade policy announcements can increase price volatility in global food commodity markets

The study investigates whether and how announcements of national trade policy changes affect global price volatility in staple food commodity markets, focusing on wheat and maize. In the context of recent shocks such as the 2022 war in Ukraine and earlier crises (2007–2008, 2010–2011), governments often use trade measures to stabilize domestic prices. Yet there is concern that such actions—especially export restrictions—can heighten global price volatility, harming other countries and undermining domestic stabilization as volatility transmits back to local markets. The authors hypothesize that policy announcements that decrease world supply (export restrictions) or increase world demand (import liberalization) raise global price volatility, with effects strongest when market stocks are low. The purpose is to provide empirical evidence using high-frequency data and policy-specific coding to inform trade-offs between domestic stabilization goals and global volatility objectives embedded in the 2030 Agenda for Sustainable Development.

Prior work highlights the welfare costs of food price volatility for both consumers and producers, particularly in developing countries where hedging options are limited. Earlier literature and policy debates focused on price levels during crises, but volatility was recognized as a key policy target due to its broad adverse effects on planning, investment, and food security. Studies have discussed the role of export restrictions and, more generally, protectionist measures in influencing global prices, often concluding that wheat markets are more affected than maize. However, empirical evidence linking specific policy announcements to high-frequency volatility has been scarce. Theoretical foundations, including the competitive storage model, suggest that stockholding moderates the transmission of shocks to prices, with asymmetries depending on whether shocks are supply-reducing or demand-increasing. Some prior arguments posited that liberal import policies likely had limited impact on volatility, but this had not been tested empirically at global scope. The present study addresses these gaps by using daily volatility measures and a policy-typed event dataset.

Design: Event-study framework assessing announcement-day and short-horizon effects of trade policy changes on global food price volatility for wheat and maize (2005–2017). Volatility Measurement: Daily volatility is estimated using range-based measures from futures prices—specifically the daily high–low log price range—capturing intraday movements and improving informational efficiency over close-to-close returns. Data Sources: - Policy announcements: Media-based dataset from Reuters (Factiva), English-language articles (Jan 2005–Jul 2017). An initial 27,507 articles were screened; 1,165 were relevant; 1,737 trade policy events were identified across commodities; 556 articles for wheat/maize presented new information and were used to timestamp announcement events. Trade policy types and directions were hand-coded (tariff vs non-tariff; direction higher/lower), mapped via rules to expected world market supply/demand shocks (e.g., higher export tax → lower world supply; lower import quota → lower world demand). Events announced on the same trading day and of the same type were aggregated to a single event. - Prices: Nearby CBOT futures for wheat and maize (Datastream/Thomson Reuters). Highest and lowest daily prices used to compute ranges; sample extends to Mar 2018 to capture post-event volatility dynamics. - Stocks: Monthly US stocks-to-use ratios (USDA WASDE) used to classify market tightness due to higher frequency availability and relevance to CME-linked price dynamics. Low-stock months defined as below the 20th percentile over 2005–2017: thresholds 0.24 (wheat) and 0.07 (maize). - Trade shares robustness: Subsample of top ten exporters/importers per commodity (FAOSTAT) to test sensitivity to market concentration. Econometric Model: Conditional Autoregressive Range model with exogenous variables (CARRX). R_t = λ_t ε_t; λ_t = ω + Σ α_i R_{t−i} + Σ β_i λ_{t−i} + Σ γ_j X_{t−j}, where X comprises policy dummies equal to 1 on the announcement day (and in persistence analyses, extended windows). Separate dummies for policy types: restrictive export (negative supply), liberal import (positive demand), liberal export (positive supply), restrictive import (negative demand). Parameters estimated via maximum likelihood assuming asymptotic normality; standard errors from the Hessian. Primary estimands are γ coefficients indicating abnormal volatility on and after announcement. Persistence Analysis: Event windows extended by sequentially adding trading days (up to 30) to evaluate duration. Aggregated Effects: Complementary monthly regressions of price variance (sum of squared daily ranges) on counts of monthly policy changes by type and stock regime to approximate cumulative impacts of clustered announcements. Software: MATLAB R2019b (Windows 11).

- Announcement-day effects: • Wheat: Restrictive export policy announcements significantly increase global wheat price volatility. Under low stocks, liberal import policy announcements also significantly raise wheat volatility. No significant effects for liberal export or restrictive import announcements. • Maize: Liberal import policy announcements significantly increase maize price volatility overall. Restrictive export announcements do not show significant effects. Stock levels do not materially moderate maize results (similar coefficients in low vs high stocks). - Stock dependence: Effects are strongest and more persistent when stocks are low. For wheat, high stocks mitigate both supply-side (export restrictions) and demand-side (import liberalization) shocks; maize stocks-to-use ratios are generally lower and show limited moderating power. - Persistence: Volatility shocks peak on announcement day and persist for roughly up to 10 trading days before converging towards baseline; persistence is longer in low-stock periods. - Clustering and cumulative impacts: Monthly regressions indicate cumulative effects when multiple announcements occur within a month. Example for wheat: one liberal import announcement is estimated to raise monthly price variance by ~45%; three such announcements raise variance by ~135%. For maize, monthly effects of liberal import policies are directionally similar but not statistically significant, likely due to fewer months with multiple events. - Robustness: Results are qualitatively similar when restricting to top exporters/importers; wheat effects are slightly stronger in this concentrated market subsample. - Null effects: No significant volatility increases detected for liberal export or restrictive import announcements, consistent with these being positive supply and negative demand shocks that stocks can more effectively buffer.

The findings support the hypothesis that policy announcements which reduce world supply (export restrictions) or increase world demand (import liberalization) raise global food price volatility, especially in tight markets with low stocks. These effects materialize immediately upon information release and can last for about ten trading days, implying that expectations formation and uncertainty drive short-run volatility around policy news. The asymmetric role of stocks is consistent with competitive storage theory: existing inventories better absorb positive supply or negative demand shocks than negative supply or positive demand shocks. Policy relevance is substantial: measures intended to insulate domestic markets can elevate global volatility, potentially feeding back into domestic prices and undermining stabilization goals. The wheat–maize contrast aligns with prior evidence that wheat markets are more sensitive to policy shocks and that maize market structure (e.g., dominance of a major exporter that generally avoids restrictions) limits the global impact of others’ restrictions. The results broaden the literature by demonstrating that import liberalization can also heighten volatility, contrary to earlier conjectures. Overall, policymakers should weigh domestic benefits against global volatility costs and consider inventory policies to dampen adverse effects.

This study provides high-frequency, policy-type-specific evidence that announcements of restrictive export and liberal import measures can increase global price volatility in wheat and maize futures, with effects most pronounced and persistent in periods of low stocks. Wheat volatility responds strongly to export restrictions and, under low stocks, to import liberalization; maize volatility increases with import liberalization. Effects persist up to about ten trading days and can cumulate when announcements cluster. Policy implications include: avoiding export restrictions and import liberalization when stocks are low; integrating potential international spillovers into policy design; and using higher stock levels as a tool to mitigate volatility effects and shorten persistence. Future research should examine implementation-day effects versus announcement effects, how timing, duration, and stringency of measures shape volatility responses, cross-commodity spillovers, the role of speculative activity, and the use of machine learning to capture moderating factors from richer announcement and contextual data.

- Focus on announcement-day effects: Implementation-day effects are not analysed; discrepancies may arise if implementation is uncertain or details emerge later. - Stock data proxy: Uses US monthly stocks-to-use ratios as a proxy for market tightness due to data frequency constraints; global stocks are only available annually. - Potential endogeneity: While same-day reverse causality is unlikely, broader lead–lag endogeneity between volatility and policy changes cannot be fully ruled out. - Data coverage: Reliance on English-language Reuters (Factiva) may underrepresent smaller countries or non-English sources; events are more likely captured for larger, more visible traders. - Market proxy: Futures prices (CBOT) are used as a global proxy; spot market volatility is unobserved, and rice is excluded due to representativeness issues. - Model scope: CARRX is optimized for event-level effects and less suited for estimating cumulative impacts from clustered announcements; monthly regressions provide only back-of-the-envelope aggregation. - Aggregation of simultaneous events: Events of the same type on the same trading day are aggregated, potentially masking heterogeneity in policy stringency. - Mechanisms: The study does not model determinants of policy changes themselves. - Weighting by trade shares: No direct weighting by country trade shares due to media-based sampling; addressed partially via robustness checks for major traders.