Implications of war on the food, beverage, and tobacco industry in South Korea

The Russia-Ukraine war, which began on February 24, 2022, disrupted global interdependencies and supply chains already strained by the Covid-19 pandemic. Sanctions on Russia and logistical disruptions contributed to price surges in key commodities, notably grains and energy, elevating inflationary pressures worldwide. Given Russia and Ukraine’s crucial roles in global wheat, corn, and soybean markets, South Korea—highly import-dependent for food—faces heightened vulnerability in its F&B sector. This study investigates how the war affects South Korea’s F&B sector stock performance by analyzing KOSDAQ F&B index returns. Addressing a gap in the literature (no prior empirical study focused on South Korea’s F&B sector under the Russia-Ukraine war), the paper aims to provide evidence useful to investors and policymakers and to forecast sectoral returns using a time-series approach.

Prior work has examined macroeconomic and geopolitical shocks on financial markets and economic growth, including effects of Covid-19 and conflicts on stock markets and commodities (e.g., Behnassi and El Haiba, 2022; Izzeldin et al., 2023; Kumari et al., 2023; Hudson and Urquhart, 2022). Studies also address determinants of stock returns and the use of statistical and machine learning methods for forecasting (e.g., He et al., 2022; Benvenuto et al., 2020; Khoa and Huynh, 2022). However, there is a lack of empirical research on the Russia-Ukraine war’s impact specifically on South Korea’s F&B sector. This paper contributes by focusing on KOSDAQ F&B returns and applying ARIMA forecasting in the context of conflict-induced supply chain disruptions.

Data: Daily stock returns (%) for the KOSDAQ F&B industry from January 1999 to October 2022 (8,491 observations) sourced from the DataGuide database. Additional descriptive and sectoral context data (e.g., net sales) were obtained from Total Solutions 2000; macro indicators from sources such as the Bank of Korea and Korea Customs Service. Baseline: A linear regression base model was first fit to provide a benchmark. Baseline error metrics: MAE = 0.13, MAPE = 13.44%, MDAPE = 11.13%. ARIMA modeling: Univariate time series forecasting using the Box–Jenkins ARIMA (p,d,q) framework. Steps: - Stationarity check using KPSS unit root test. - Identification of p and q via ACF and PACF; PACF showed significance around lag 3. - Model order selection guided by AIC and SIC. - Data split reported as 70:30; 8,491 observations were divided into 5,943 test and 2,548 training observations; elsewhere the training span is described as Jan 1999–Dec 2021, with forecasts compared to actuals thereafter. - Residual diagnostics: residual plots showed near-zero mean and approximately uniform variance; density plots suggested acceptable error distribution. Selected model: ARIMA(2,2,3). Forecast accuracy for ARIMA: RMSE = 0.012, MAE = 0.007, MAPE = 0.015. Figures show actual vs. forecasted returns and residual diagnostics. The ARIMA model outperformed the baseline regression in the reported metrics.

- The Russia-Ukraine war is associated with a negative trend in South Korea’s KOSDAQ F&B sector returns for several months; sector returns declined as the conflict intensified. - The ARIMA(2,2,3) model achieved strong predictive performance (RMSE = 0.012; MAE = 0.007; MAPE = 0.015), outperforming a linear regression baseline (MAE 0.13; MAPE 13.44%; MDAPE 11.13%). - Descriptive statistics indicate high volatility (standard deviation ≈ 1.886) and leptokurtic, negatively skewed return distribution (kurtosis ≈ 4.97; skewness ≈ −0.385). - Macro and trade indicators reinforce vulnerability: grain import prices per ton in February 2022 were up 26% year over year and 47.4% versus February 2020; the CPI for food rose by more than 5% since April 2022; South Korea’s food self-sufficiency has declined (overall ~45.8% in 2019; grain self-sufficiency ~21%), reflecting heavy import dependence. - Despite a longer-term increasing trend in net sales for the F&B industry domestically, monthly F&B stock returns showed a downward shift around February 2022 concurrent with the onset of the war.

The findings suggest that conflict-driven shocks to global food supply chains and commodity prices transmit quickly to sectoral equity performance in an import-dependent economy like South Korea. The observed decline in KOSDAQ F&B returns during the war months addresses the study’s core question about the war’s economic implications for the sector. Reliable ARIMA-based forecasts can aid investors in risk management and timing decisions under elevated uncertainty. For policymakers, the results underscore the need to mitigate exposure to external supply disruptions by strengthening domestic agribusiness capacity, logistics resilience, and smart-agriculture initiatives. Enhancing self-sufficiency in key agri-food inputs, fostering ICT-enabled farming, and streamlining supply chains could reduce vulnerability to future geopolitical and logistics shocks while supporting stable sectoral performance.

The Russia-Ukraine war adversely impacted South Korea’s F&B sector stock returns, reflecting heightened input costs, supply chain bottlenecks, and inflationary pressures. An ARIMA(2,2,3) model provided accurate short-term forecasts, evidencing a negative return trend during the conflict months. To bolster food security and sector resilience, South Korea should undertake urgent and long-term reforms across agriculture, logistics, and industry, including smart agriculture investments and policies that enhance domestic production capacity. Given South Korea’s technological and infrastructural strengths, progress toward greater self-reliance is feasible and aligns with broader sustainability objectives. Future research should extend the analysis across countries and incorporate energy price dynamics to better capture intertwined food and energy security risks.

The analysis focuses on South Korea and relies on a univariate time series of sectoral returns, without explicitly modeling macroeconomic covariates or cross-country linkages. Data constraints limit granularity on firm-level exposures and sub-sector heterogeneity. The training/testing split reporting is inconsistent across sections. Future work should include multi-country comparisons, integrate energy price variables and other macro drivers, and explore multivariate or structural models to identify channels and improve generalizability.