Analysis of social welfare impact of crop pest and disease damages due to climate change: a case study of dried red peppers

The study examines how climate change affects agricultural production and social welfare through direct climatic effects and indirect effects via increased outbreaks of pests and diseases (P&D). Motivated by accelerating global warming and changing precipitation patterns, particularly in East Asia, the authors focus on dried red peppers in Korea, a staple seasoning crop with large cultivation area and notable vulnerability to climatic variability and P&Ds. The research question is to quantify how much social welfare changes under climate change when accounting for both direct climatic effects on yield and indirect effects through P&D damages, and to compare welfare outcomes with and without P&D impacts. The purpose is to provide a comprehensive, monetized measure of climate change impacts on a specific crop's market, highlighting implications for food security and the need for pest management.

Prior research documents direct climate impacts on crop productivity (e.g., temperature and precipitation effects on rice, grains, and vegetables) and indirect impacts through increased P&D and weeds under warming and altered moisture regimes. Studies in Korea and elsewhere have shown specific climatic drivers of P&D outbreaks in rice and ginseng, and broader evidence that rising CO2, temperature, and water stress alter pest population dynamics. Economic studies have valued welfare impacts of disease outbreaks and pest control in livestock and crops using equilibrium models and ecological-economic approaches. However, few studies jointly analyze both direct and indirect (P&D-mediated) climate effects on crop yields and translate them into social welfare changes for a given crop. This work addresses that gap by integrating P&D damage modeling, yield modeling, and an equilibrium displacement model (EDM) to monetize welfare impacts under climate scenarios.

Study materials and data: The focal crop is dried red pepper in Korea, grown in open fields and economically significant (largest area share among open-field vegetables). Four major P&Ds are considered: phytophthora blight (PB), anthracnose (Ath), viral diseases (VD), and tobacco moths (TM). P&D damage data come from the National Crop Pest Management System (NCPMS) for 2014–2021, collected eight times per year at 15-day intervals between June 1 and September 16 across 59 cities (averaged from 1,657 observation sites). Damage rates are measured as damaged heads rate (PB, VD) and damaged fruits rate (Ath, TM). Climate variables (temperature, rainfall, humidity, sunshine hours) are from the Korea Meteorological Administration (KMA) and spatially interpolated in QGIS to match observation sites. Yield data (kg/10a) by province (8 provinces) for 2014–2021 are sourced from Statistics Korea (KOSTAT). Climatic variables are aggregated by key crop growth periods: planting (April), growing (May–July), and harvesting (August–October). Deviations from long-term averages (1970–2000) for temperature and humidity are included to capture abnormality. Empirical models: 1) P&D damage model: A left-censored panel Tobit model with random effects and Chamberlain random effects (CRE) correction is used due to many zero damage-rate observations and short panel length. Dependent variable: P&D damage rate for each P&D (PB, Ath, VD, TM). Regressors: city latitude/longitude, contemporaneous climatic factors (temperature, rainfall, humidity, sunshine hours), and deviations from 1970–2000 averages for temperature and humidity to control fixed effects and abnormality. Panel indices: cities i=1..59, sessions t=1..64 (2014–2021, eight sessions per year). 2) Yield model: A fixed-effects panel regression at the provincial level (8 provinces, 2014–2021) evaluates direct and indirect climate effects on yield (kg/10a). Dependent variable: yield. Regressors: direct climatic variables by period (average temperature, rainfall, sunshine hours during planting/growing/harvesting), deviations from historical temperature averages by period, and predicted P&D damage variables (from the P&D model) capturing indirect effects. To avoid multicollinearity, Ath and TM predicted values are combined (ATM) due to high correlation (r≈0.68). Fixed effects are chosen based on tests (Hausman/rejection of random effects via Breusch-Pagan LM result). Multicollinearity checks showed acceptable VIF (avg 4.83). 3) Equilibrium displacement model (EDM): A comparative static framework quantifies how exogenous shifts in supply due to temperature changes (direct and P&D-mediated) affect equilibrium price (P) and quantity (Q), and then social welfare. Assumptions: initial price P0=9,140 KRW/kg (KREI OASIS, 2014–2021), initial quantity Q0=74,672 metric tons (KOSTAT, 2014–2021). Supply elasticity εP∈[0,0.2] (drawn from related crops: subtropical vegetables, rice, beans/tofu) and demand elasticity ηP∈[-0.24,-0.16] (vegetable demand literature). Demand shock U=0. RCP 6.0 temperature change scenarios for 2030, 2040, 2050 are used, with period-specific temperature changes: planting 12.1, 13.8, 14.1°C; growing 22.4, 22.3, 24.2°C; harvesting 21.3, 21.9, 21.3°C, respectively (relative to present averages). Yield changes under each scenario are computed from the yield model, both with and without P&D damages, to construct the exogenous supply shift K. EDM solves for changes in equilibrium price and quantity and computes social welfare changes (areas under demand and above supply), with sensitivity analyses across elasticity values. Key parameter tables: Initial values and elasticity ranges (Table 4); RCP period-specific temperature changes (Table 5).

P&D damage model (Table 6): - PB: Damage increases with temperature deviation from the 1971–2000 average (Avgt dev, positive, significant), and decreases with higher sunshine hours (Sun, negative, significant) and humidity deviation (Avgh dev, negative, significant). - Ath: Significant effects of average temperature (Avgt, negative) and its square (Avgt^2, positive), average humidity (Avgh, positive), and deviations in temperature and humidity (Avgt dev positive; Avgh dev positive). The damage rate has a U-shaped relation with temperature, with a minimum at approximately 30.3°C. - VD and TM: Both show significant effects of Avgt (negative) and Avgt^2 (positive), indicating U-shaped relationships with minima at about 37.5°C (VD) and 28.6°C (TM). Average humidity (Avgh) increases damage, sunshine hours reduce it, and temperature deviation (Avgt dev) increases damage; humidity deviation (Avgh dev) reduces damage. Yield model (Table 7): - Direct climatic effects: Temperature in planting and harvesting periods positively affects yield; growing-period temperature negatively affects yield (signs align with seasonal agronomy). Quantitatively, a 1°C increase in planting temperature raises yield by about 25.3 kg/10a; a 1°C increase in harvesting temperature raises yield by about 47.7 kg/10a. Rainfall effects are period-specific: +10 mm in planting increases yield by ~2.5 kg/10a, while +10 mm in growing and harvesting reduces yield by ~2.5 kg and ~2.8 kg/10a, respectively. - Indirect effects via P&D: Predicted P&D damage variables have negative coefficients (e.g., PB ≈ -245 kg/10a; ATM and VD negative), indicating P&D damages suppress yields, offsetting positive direct temperature effects. Yield under RCP 6.0 scenarios (Table 8): - Relative to the present (241.8 kg/10a), projected yields without P&D damages rise to 276.3 (2030), 286.0 (2040), 302.5 (2050) kg/10a. With P&D damages, yields are lower: 275.3 (2030), 268.0 (2040), 278.8 (2050) kg/10a. The yield gap (without minus with P&D) widens over time: 1.0 (2030), 18.1 (2040), 23.8 (2050) kg/10a. EDM social welfare impacts (Tables 9–10): - Holding supply elasticity at 0.1 and varying demand elasticity: With P&D damages vs without, social welfare losses are larger by approximately 9.3, 161.5, 239.2 billion KRW in 2030, 2040, 2050 for ηP=-0.16; 7.1, 125.8, 183.0 billion KRW for ηP=-0.20; and 9.8, 174.7, 251.2 billion KRW for ηP=-0.24. - Holding demand elasticity at -0.20 and varying supply elasticity: Additional social welfare losses with P&D damages are about 6.8, 116.1, 176.4 billion KRW (εP=0.0) in 2030, 2040, 2050; 7.3, 130.4, 186.1 (εP=0.1); and 7.6, 138.2, 191.4 (εP=0.2) billion KRW. - Overall range: The difference in social welfare with vs without P&D damages across scenarios and elasticities ranges from 6.8 to 251.2 billion KRW, equivalent to roughly 1%–36.2% of the 2021 dried red pepper production value (694.2 billion KRW, 2015=100). Interpretation: While warming alone tends to increase yields and social welfare, indirect P&D damages increasingly erode these gains over time, leading to higher prices and lower quantities than in a counterfactual without P&D impacts.

The study’s sequential framework shows that climate change influences P&D dynamics, which in turn negatively affect yields and market outcomes. The P&D model evidence that temperature deviations elevate damage rates, and that Ath, VD, and TM exhibit U-shaped temperature relationships with minima at specific temperatures, implies that increasingly abnormal climatic conditions will intensify P&D pressures beyond agronomically favorable temperature ranges. The yield model demonstrates that even though moderate warming during planting and harvesting can boost yields, these gains are offset by P&D damages and by adverse rainfall timing effects during growing and harvesting periods. Integrating these into the EDM reveals that supply shifts are smaller (or even leftward relative to the no-P&D case), raising prices and reducing quantities, thereby lowering social welfare relative to a counterfactual without P&D impacts. The difference in welfare magnifies over time and is sensitive to demand and supply elasticities, underscoring that market responsiveness shapes the magnitude of welfare losses. These findings directly address the research question by quantifying, in monetary terms, the sizable indirect costs of climate change via P&D on a key Korean crop and highlight the broader relevance for food security as climate-induced pest pressures scale globally.

The paper provides an integrated, monetized assessment of climate change impacts on the dried red pepper market by linking climate-driven P&D damages to yields and market welfare. Key contributions include: (1) a panel Tobit analysis of P&D damage responses to climatic factors, (2) a fixed-effects yield model capturing both direct climatic and indirect P&D effects across growth periods, and (3) an EDM translating supply shifts under RCP 6.0 scenarios into price, quantity, and social welfare changes. Results indicate that while rising temperatures can increase yields and welfare, P&D damages increasingly suppress these gains, generating substantial welfare losses (up to 251.2 billion KRW by 2050 under plausible elasticities). Policy implications include the need for rigorous pest prediction and control, improved water management aligned with critical growth stages, and resilience strategies in open-field vegetable systems. Future research should incorporate demand-side shocks, consumer responses, and government price stabilization policies into the EDM; extend to multi-crop systems and trade linkages; and employ longer panels and high-resolution pest-climate projections to refine forecasts.

- The EDM considers supply-side shocks only (U=0), omitting potential demand-side changes (e.g., consumer responses, substitution, income effects) and government policy interventions that could shift demand or stabilize prices. - Forecast uncertainty increases with the time horizon; projections through 2050 are subject to greater error due to the limited temporal scope of the historical data (2014–2021). - Elasticity parameters are drawn from related crops/literature rather than crop-specific estimates for Korean dried red peppers, introducing parameter uncertainty. - P&D and climate data are seasonal and averaged across sites/provinces; unobserved heterogeneity and spatial dynamics may persist despite CRE and fixed-effects controls. - The study focuses on temperature-driven scenarios (RCP 6.0) and does not fully simulate other climate dimensions (e.g., extreme events, precipitation variability) in the EDM stage.