# Executive Summary: Commodity Volatility and Cross-Commodity Linkages

Data coverage: 71 commodities with 796 monthly observations between years 1960 and 2026.

Volatility and drawdowns:
- Mean annualized volatility across commodities: 0.232
- Maximum annualized volatility: 0.464 (most volatile commodity: BANANA_US (Banana, US, Agriculture))
- Minimum annualized volatility: 0.075 (least volatile commodity: TOBAC_US (Tobacco, US import u.v., Soft Commodities))
- Worst maximum drawdown magnitude: 0.950 (commodity: SUGAR_WLD (Sugar, world, Soft Commodities))

Group-level volatility (mean volatility by group):
- Energy: mean 0.270, median 0.268, n=17
- Precious Metals: mean 0.198, median 0.187, n=3
- Industrial Metals: mean 0.211, median 0.222, n=7
- Agriculture: mean 0.204, median 0.195, n=26
- Soft Commodities: mean 0.234, median 0.223, n=14
- Fertilizers: mean 0.313, median 0.295, n=4

Correlation regimes and contagion:
- Full-sample average absolute correlation: 0.111
- Full Sample average |corr|: 0.111
- Pre-2000 average |corr|: NaN
- 2000-2008 average |corr|: 0.210
- 2009-2019 average |corr|: 0.135
- 2020 onwards average |corr|: 0.174
- Rolling 36-month average |corr|: 0.196, average network density: 0.045

Volatility clustering and regimes:
- Number of commodities with GARCH-type models estimated: 0
- Realized 36-month, EWMA, and GARCH volatility estimates broadly confirm strong volatility clustering and persistent regimes across key commodity benchmarks.

Network spillovers:
- High-correlation windows are associated with dense spillover networks and elevated co-movement across commodities, indicating periods of systemic stress.