Research insight · Issue 008
When commodity shocks stop staying local
Commodity stress becomes systemic when shocks cross commodity groups.
- Published
- Data through
- 11 September 2026
ContentsExecutive summary
Executive summary
Commodity stress becomes systemic when shocks cross commodity groups.
- 01Cross-group spillovers rise from 32.2% to 49.4%.
- 02The largest increases appear across metals and energy links.
- 03Energy shifts towards a net transmitting role.
- Low regime
- 32.2%
- High regime
- 49.4%
- Increase
- +17.2 pp
- Current
- 55.3%
- Peak
- 65.2%
Why it matters
Higher connectedness means shocks travel across more commodity groups, creating a broader transmission system.
Analysis 01
When commodity connectedness rises
Commodity connectedness changes materially through time. The index ranges from 33.4% at its 2014 low to 65.2% at the 2022 peak, compared with a long-run median of 47.1%. The latest reading is 55.3%, leaving the system below its previous peak but materially above its historical median. The result establishes that commodity transmission is strongly state-dependent. Some periods remain relatively segmented, while others move into a substantially more integrated system in which shocks have more routes through which to propagate.
Total connectedness reaches a low of 33.4% on 31 March 2014 and a peak of 65.2% on 19 July 2022. The median is 47.1% and the latest reading is 55.3%.
Analysis 02
The source of commodity shocks rotates
Commodity transmission is highly uneven. Corn is currently the strongest net transmitter at +0.153, followed by silver, soybeans and Brent crude. At the other end, sugar is the largest net receiver at −0.209, while natural gas and copper also absorb more spillovers than they send. The result shows that systemic importance is not fixed to one commodity group. Energy does not automatically dominate transmission, and the commodities carrying shocks can change materially with the state of the network. The source of commodity-system risk rotates through time.
Corn is the largest net transmitter at +0.153 and sugar the largest net receiver at −0.209. Silver, soybeans, Brent crude, platinum, WTI crude and gold are net transmitters; wheat, copper and natural gas are net receivers.
Analysis 03
The network becomes less contained
When commodity connectedness is low, 32.2% of total spillovers cross commodity groups. In the high-connectedness regime, that rises to 49.4%, an increase of 17.2 percentage points. The result shows that higher connectedness is not simply stronger transmission within existing commodity families. A much larger share of shocks begins crossing the boundaries between them.
Cross-group spillovers rise from 32.2% in the low-connectedness regime to 49.4% in the high-connectedness regime, an increase of 17.2 percentage points.
Analysis 04
The transmission map broadens under stress
In low-connectedness periods, cross-group transmission remains relatively contained, with the strongest links concentrated around the metals channel. In high-connectedness regimes, the map broadens materially. Energy, agriculture and industrial metals all become more strongly linked, while precious metals remain an important source of transmission. The important point is that higher connectedness changes the architecture of the system, not just the strength of individual links. Stress turns a more segmented commodity network into a broader transmission system.
Columns identify shock transmitters and rows identify receivers. Precious metals to industrial metals rises from 0.059 to 0.096; energy to industrial metals rises from 0.022 to 0.052 and energy to agriculture from 0.014 to 0.038. Diagonal own-group cells are excluded.
Analysis 05
Stress opens new routes
The largest changes occur across the metals channel. Precious metals → industrial metals rises by 0.037, while the reverse direction increases by 0.036. Energy also becomes more important. Transmission from energy → industrial metals rises by 0.030 and energy → agriculture by 0.024. The result shows that high connectedness does more than strengthen existing relationships. It creates a broader set of economically important routes through which disturbances can spread. Commodity stress becomes systemic as shocks gain more paths across the network.
The largest increases are precious metals to industrial metals (+0.037), industrial metals to precious metals (+0.036), energy to industrial metals (+0.030) and energy to agriculture (+0.024).
Analysis 06
Energy changes role as connectedness rises
Precious metals remain a net source of cross-group spillovers in both regimes. Energy behaves differently. It moves from a slight net receiver in the low-connectedness regime to a net transmitter when connectedness is high. Agriculture becomes a somewhat larger net receiver, while industrial metals remain the strongest receiving group. The result shows that rising connectedness changes not only how far shocks travel, but also which commodity groups send them and which absorb them. The direction of systemic transmission changes with the regime.
Energy shifts from −0.001 in the low regime to +0.004 in the high regime. Precious metals remains positive (+0.009 to +0.008), while agriculture (−0.001 to −0.003) and industrial metals (−0.007 to −0.009) remain net receivers.
Methodology
How the analysis is constructed
- Sample period
- 04 January 2012–11 September 2026
- Coverage
- 11 exchange-traded commodity exposures across four groups
- Data source
- Yahoo Finance, accessed through yfinance. Daily adjusted closing prices are converted into log returns.
- Data through
- 11 September 2026
- Commodity exposures
- Energy — USO, BNO, UNG
- Agriculture — CORN, WEAT, SOYB, CANE
- Industrial metals — CPER
- Precious metals — GLD, SLV, PPLT
Data
The common sample retains dates with observations available across the full universe.
Rolling estimation
Connectedness is estimated over a 120-observation rolling window. Each window uses a VAR(1) with no deterministic trend.
Forecast-error variance decomposition
Directional spillovers are estimated using a generalised FEVD with a 10-step forecast horizon. The decomposition is row-normalised so each receiver’s variance contributions sum to one.
Directional measures
For each commodity:
- FROM measures spillovers received from the rest of the system.
- TO measures spillovers transmitted to the rest of the system.
- NET = TO − FROM
Positive NET identifies a net transmitter. Negative NET identifies a net receiver.
Total connectedness
The total connectedness index is the average off-diagonal FEVD contribution across the system, expressed as a percentage.
Regime comparison
Rolling connectedness values are classified using the empirical distribution of the total connectedness index. The low-connectedness regime contains the bottom 20% of rolling observations. The high-connectedness regime contains the top 20%. For each regime, FEVD matrices are averaged across all qualifying rolling windows. Cross-group transmission is then aggregated across four commodity families: Energy, agriculture, industrial metals and precious metals. To avoid larger groups dominating mechanically, group-to-group transmission is normalised by the number of possible directed commodity pairs. Own-variable effects are excluded. Reported outputs include cross-group spillover shares, group-level net transmission and changes in directional transmission between regimes. The purpose is to identify whether higher connectedness changes only the strength of spillovers, or the routes through which they travel.
Research workflow
- Data
- Daily adjusted prices for 11 exchange-traded commodity exposures
- Returns
- Daily log returns
- Rolling window
- 120 observations
- Model
- VAR(1), no deterministic trend
- Spillover measure
- Generalised forecast-error variance decomposition
- FEVD horizon
- 10 steps
- Normalisation
- Row-normalised variance contributions
- Total connectedness
- Average off-diagonal FEVD contribution
- Directional measures
- TO, FROM and NET spillovers
- Commodity groups
- Energy, agriculture, industrial metals and precious metals
- Low regime
- Bottom 20% of rolling connectedness
- High regime
- Top 20% of rolling connectedness
- Group transmission
- Mean contribution per possible directed commodity pair
- Cross-group share
- Share of total spillovers occurring between commodity groups
- Regime comparison
- High-minus-low changes in directional transmission
Higher commodity connectedness means that spillovers are no longer concentrated within individual commodity groups. The rise in cross-group transmission shows that shocks increasingly move between energy, agriculture and metals rather than remaining contained within one part of the system. The regime comparison also shows that transmission roles change. Precious metals remain an important source of cross-group spillovers, while energy shifts towards a net transmitting role as connectedness rises. The result is structural rather than causal. It identifies how the routes available to shocks change with the state of the commodity network, not a single historical shock path.
Research implications
What the evidence shows
The evidence in this report shows that commodity spillover structure changes materially across connectedness regimes.
- 01Total connectedness ranges from 33.4% to 65.2% across the sample.
- 02The current reading is 55.3%, above the long-run median of 47.1%.
- 03Cross-group spillovers rise from 32.2% to 49.4% in high-connectedness regimes.
- 04Precious metals remain an important cross-group transmitter.
- 05Energy shifts from a slight net receiver to a net transmitter as connectedness rises.
- 06The strongest increases occur across metals and energy-linked transmission routes.
Together, the results suggest that commodity stress becomes more systemic when shocks stop remaining inside individual commodity groups. Monitoring connectedness alongside price moves can therefore help distinguish between a local commodity shock and one with a much broader cross-market transmission risk.
Publication record
Sources and disclosures
Data
- Yahoo Finance, accessed through yfinance.
- 11 exchange-traded commodity exposures across energy, agriculture, industrial metals and precious metals.
- Data through 11 September 2026.
Methodology reference
- Pesaran, M. H. and Shin, Y. (1998). “Generalized impulse response analysis in linear multivariate models.” Economics Letters, 58(1), 17–29.
- Diebold, F. X. and Yilmaz, K. (2012). “Better to give than to receive: Predictive directional measurement of volatility spillovers.” International Journal of Forecasting, 28(1), 57–66.
- Diebold, F. X. and Yilmaz, K. (2014). “On the network topology of variance decompositions: Measuring the connectedness of financial firms.” Journal of Econometrics, 182(1), 119–134.
Software
Python, pandas, NumPy, statsmodels, matplotlib, NetworkX and yfinance.
This publication has been prepared by LB Research for informational and educational purposes only. It does not constitute investment advice, investment research as defined under applicable regulation, a recommendation, or an offer to buy or sell any financial instrument.
The analysis is based on publicly available market data and the methodologies described within this publication. While reasonable care has been taken in preparing this report, no representation or warranty is made regarding the accuracy, completeness or timeliness of the information presented.
Any opinions expressed reflect the author's judgement at the publication date and may change without notice. Past performance is not indicative of future results.
Readers remain solely responsible for their own investment decisions.
© 2026 LB Research. All rights reserved. Questions or feedback: Contact LB Research
Download canonical PDFIn development: extending this research into cross-market structure and risk intelligence.