Volatility in Commodity Markets Part 1: How Shocks Spread

Book: Commodities: Markets, Performance, and Strategies
Editors: H. Kent Baker, Greg Filbeck, Jeffrey H. Harris
Publisher: Oxford University Press, 2018
ISBN: 9780190656010

When One Market Sneezes

Chapter 18 by Frankie Chau and Rataporn Deesomsak tackles a question that matters for every commodity investor: when volatility hits one commodity market, does it spread to others? And if so, how fast and how far?

This is not just an academic exercise. If you hold a diversified commodity portfolio and volatility spillovers are high, your diversification might not protect you as much as you think. If you are a regulator, understanding how shocks transmit between markets helps you figure out where to focus your attention during a crisis.

The Boom-Bust Cycle

To understand why this research matters, look at what happened to commodity prices between 2005 and 2016. Most commodity prices started climbing in early 2005, creating the longest commodity price boom since World War II. Oil briefly touched $150 a barrel in July 2008. Then global demand collapsed during the financial crisis and oil prices halved.

The authors identify three main drivers of this boom-bust cycle:

1. Growing demand from emerging economies, especially China
2. Low interest rates and dollar depreciation
3. Increased liquidity from institutional investors, hedge funds, and ETFs entering commodity markets

Investment fund activity in commodities was about $330 billion by early 2012, roughly nine times higher than in the early 2000s. That is a lot of new money chasing returns in markets that were historically dominated by producers and consumers.

What the Existing Research Says

Before building their own model, Chau and Deesomsak review what prior studies found about volatility spillovers.

Olson, Vivian, and Wohar (2014) looked at energy and equity markets. They found that low S&P 500 returns caused big increases in energy index volatility. But the reverse relationship was weak. Stock market stress spills into energy, but energy stress does not spill much into stocks.

Antonakakis and Kizys (2015) studied commodity and currency markets from 1987 to 2014. They found that volatility spillovers from gold and silver reached unprecedented levels during the financial crisis. But after the crisis passed, gold and silver weakened as net transmitters of shocks.

Nazlioglu, Erdem, and Soytas (2013) focused on energy and agricultural markets. Before the food price crisis, no significant risk transmission existed between oil and agricultural commodity markets. After the crisis, oil market volatility started spilling into agricultural markets. The crisis fundamentally changed the relationship between these markets.

Yaya, Tumala, and Udomboso (2016) studied oil and gold specifically. They found volatility spillovers between these two markets from 1986 to 2015. But the direction of return spillovers shifted. Before the financial crisis, spillovers went both ways. After the crisis, the flow became one-directional: from gold to oil.

The common thread in all of these studies is that the financial crisis of 2007-2008 was a structural break. The way commodity markets interact with each other and with financial markets changed after the crisis.

The Gap in the Research

Most prior studies focused on pairs or small groups: energy and equities, energy and metals, energy and agriculture. Very few looked at how all three major commodity classes interact with each other simultaneously.

Chau and Deesomsak set out to fill that gap. They examine nine commodity futures across three categories:

• Agricultural: corn, coffee, and soybeans (from CBOT and ICE)
• Energy: Brent crude oil, WTI crude oil, and natural gas (from NYMEX)
• Metals: copper, gold, and silver (from COMEX)

Their data spans from April 1990 to December 2016, which is a much longer window than most prior studies. This matters because it captures the early 1990s global recession, the dot-com bubble, the commodity boom of the 2000s, and the financial crisis and its aftermath.

The Approach: Diebold-Yilmaz Spillover Index

The methodology comes from Diebold and Yilmaz (2012). Without getting too deep into the math, the basic idea is:

1. Estimate a vector autoregression (VAR) model using the volatility of all nine commodity futures
2. Decompose the forecast error variance to figure out how much of each market’s volatility comes from its own shocks versus shocks from other markets
3. Aggregate these results into a total spillover index and directional spillover measures

The total spillover index tells you what percentage of total volatility across all nine markets comes from cross-market spillovers rather than each market’s own shocks.

The directional spillovers tell you which markets are net transmitters of volatility (they send more shocks than they receive) and which are net receivers (they absorb more shocks than they send).

The authors call their index the Commodity Volatility Spillover Index (CVSI). They compute it both as a static full-sample average and as a time-varying measure using rolling windows.

Initial Look at the Data

Some basic facts from the descriptive statistics:

• Energy prices are the most volatile. Natural gas has a weekly standard deviation of 7.49 percent, far above any other commodity in the sample.
• Almost all commodity futures returns are negatively skewed and have fat tails (high kurtosis). This means extreme negative moves are more common than extreme positive moves, and extreme moves in general happen more often than a normal distribution would predict.
• The highest correlations are within sectors: corn and soybeans at 0.61, gold and silver at 0.74, and Brent crude and WTI at 0.89.
• Cross-sector correlations are much lower but still statistically significant. This suggests some interconnection between commodity classes, but the strength varies.

Coffee is the odd one out. It has low correlations with almost everything else, including other agricultural commodities.

Why This Framework Matters

The beauty of the Diebold-Yilmaz approach is that it goes beyond simple correlations. Two markets might have low correlation in normal times but high volatility spillovers during crises. Simple correlation cannot capture this asymmetry. The spillover index framework can.

It also distinguishes between directional flows. Knowing that crude oil transmits volatility to agricultural markets, but not the other way around, is far more useful than just knowing the two are “correlated.”

For investors, this matters because it affects portfolio construction. If you think you are diversified across energy, metals, and agriculture, but all three are receiving volatility shocks from the same source (say, crude oil), your diversification is weaker than it appears.

For regulators, the directional information helps identify systemically important markets. If one commodity is the primary net transmitter of volatility across all others, that is where regulatory attention should focus during a crisis.

What’s Coming in Part 2

In the next post, we will look at the actual empirical results. Which commodity markets transmit the most volatility? Which ones absorb it? How did the financial crisis change the spillover dynamics? And what does the time-varying analysis reveal about how quickly volatility can spread across commodity markets?

Spoiler: the results challenge some conventional wisdom about where risk really comes from in commodity markets.

Previous: How to Benchmark Commodity Performance (And Why It’s Tricky)

Next: Volatility in Commodity Markets Part 2: Cross-Market Evidence