作者
蔡釗旻* (中央銀行經濟研究處)

Author
Chao-Min Tsai^* (Department of Economic Research, Central Bank of the Republic of China (Taiwan))

中文摘要
本文以 Adrian et al. (2019) 與 Prasad et al. (2019) 提出的經濟成長風險值 (Growth-at-Risk, GaR) 為基礎，採用固定效果追蹤資料的分量迴歸模型 (Machado and Santos Silva, 2019)，探討氣候實體風險對未來 3 季至 12 季經濟成長分配的影響。本文建構包含臺灣在內之 43 個主要經濟體，2000 年第 1 季至 2025 年第 1 季的追蹤資料，並以地表平均氣溫之月變動值計算的溫差平均值、偏態與峰態，分別衡量氣候變動、方向性風險與尾端風險。 實證結果顯示，第一，溫差平均值對短期經濟成長影響有限， 惟於較長的預測區間，氣溫持續攀升會加劇低分量經濟成長的下行風險；第二，溫差偏態對各分量下的經濟成長率多呈顯著的負向效果， 顯示氣候變遷的方向性風險會削弱經濟成長動能；第三，溫差峰態則主要影響經濟成長動能疲弱期間之下行風險；第四，情境分析進一步指出，氣候實體風險對經濟成長的影響具有累積性，且低分量經濟成長率對溫差平均值衝擊更為敏感；此外，無論異常高溫或異常低溫方向之風險提高，均會使未來 9 季與 12 季經濟成長率逐步惡化，且其負向效果在高分量下更為明顯，反映該風險存在累積性與不對稱性。 本文結果顯示，僅分析平均經濟成長效果可能低估氣候實體風險，政策評估宜納入經濟成長分配、氣候衝擊的持續性效果與尾端風險。

Abstract
This research examines the effects of physical climate risks on the distribution of economic growth over the next 3 to 12 quarters. Building on the Growth-at-Risk (GaR) framework of Prasad et al. (2019), it employs a fixed-effect panel quantile regression model following Machado and Santos Silva (2019). The analysis uses quarterly panel data for 43 major economies, including Taiwan, from 2000Q1 to 2025Q1. Physical climate risks are measured by the mean, skewness, and kurtosis of monthly changes in surface temperature, which capture climate change, directional risk, and tail risk, respectively. The empirical results show, first, that the mean temperature change has limited effects on short-term economic growth, but over longer forecast horizons, persistent temperature increases intensify downside risks to growth, especially at lower quantiles. Second, temperature skewness has a significantly negative effect on economic growth across most quantiles, suggesting that directional climate risks weaken growth momentum. Third, temperature kurtosis mainly influences downside risks when growth momentum is weak. Fourth, scenario analysis indicates that the impact of physical climate risks on growth is cumulative, and lower-quantile growth is more sensitive to shocks in mean temperature changes. Finally, higher risks of either extreme heat or extreme cold gradually worsen economic growth over the next 9 and 12 quarters with stronger negative effects at higher quantiles, highlighting the cumulative and asymmetric nature of directional climate risks. This paper suggests that focusing only on average growth effects may underestimate physical climate risks. Policymakers should therefore account for the full distribution of economic growth, the persistence of climate shocks, and tail risks.

中文關鍵字
氣候實體風險、溫差平均值、溫差偏態、溫差峰態、情境分 析、金融情勢、總體金融脆弱性、外部不確定性

Keywords
Physical Climate Risk, Mean of Temperature Changes, Skewness of Temperature Changes, Kurtosis of Temperature Changes, Scenario Analysis, Financial Conditions, Macro-Financial Vulnerability, External Uncertainty