July 29, 2008 - Crude Oil and Natural Gas Prices Reliably Intertwined? (Revised 7/30/08 to Append Reader Comments)

Reader Henry Bee of Vancouver, Canada asks:

"You have probably heard of the historical 6:1 crude oil/natural gas price ratio. This relationship is said to be mean reverting based on the thermal equivalence of the two commodities. Does this ratio have any predictive power for the future prices of oil or natural gas? If there is no predictive power for this ratio, then it could mean that the thermal equivalence itself shifts over time. And hedge funds who are long natural gas right now are making a huge fundamental mistake."

If there are relationships, we hypothesize that a high (low) crude oil-natural gas price ratio should predict future changes in the prices of natural gas of crude oil to decrease (increase) the ratio. Using the monthly composite U.S. refiner cost of crude oil (nominal dollars per barrel) and the monthly U.S. wellhead natural gas price (nominal dollars per thousand cubic feet) for January 1976 through April 2008 (388 months), we find that...

The following chart plots the crude oil-natural gas price ratio over the entire sample period. Visual inspection reveals no reversion level and no regular cycles. The range of the ratio is roughly 5 to 20.

For precision, we try scatter plots to relate the ratio to future price changes in the commodities.

The following scatter plot relates the next-month change in U.S. wellhead natural gas price to current month crude oil-natural gas price ratio over the entire sample period. The Pearson correlation for the two series is 0.15 and the R-squared statistic is 0.02, indicating that the ratio explains about 2% of the variation in natural gas price the next month. Extending the forecast horizon to three (six) months yields an R-squared statistic of 0.05 (0.05). Although the price ratio explains little of the future price change, it does appear that high levels of the ratio may suppress volatility in natural gas price.

A similar analysis of the next-month change in crude oil price yields a much weaker Pearson correlation of 0.02 and an R-squared statistic of 0.00, indicating no predictive power.

Why might a stronger relationship not exist?

Possible explanations are:

1. Because of differences in transport logistics, crude oil is largely a global resource while natural gas is largely continental (in this case, North America). International economics and politics may therefore perturb crude oil price to a much larger degree.
2. Costs of storage may be different for oil and gas, thereby making one more sensitive than the other to oversupply.
3. Investments in fixed assets to utilize one or the other may retard market reaction to a price mismatch.

There may be a way to play the mismatch, but a set ratio does not appear to be it.

In summary, simple analyses indicate that the crude oil-natural gas price ratio has little (no) predictive power for natural gas (crude oil) price at forecast horizons of one to a few months.

For related research, see Blog Synthesis: Investing/Trading in Commodities and Commodity Futures.

Reader Bob Tait offers the following comments:

"A better interpretation would, to my mind, incorporate the shifting relative demand picture for oil and gas in the U.S. In the period from the middle 1970s through the early 1980s, natural gas was in low demand and therefore relatively cheap. In the late 1970s through the early 1990s, both commodity chemical producers and power utilities invested in plant and equipment to take advantage of the cheap, readily available natural gas. By the mid-1990s, demand for natural gas in the U.S. was relatively high even as global demand for oil was flat. By the turn of the century, global demand for oil began to increase faster than the U.S. demand for natural gas. Consequently, oil became relatively more expensive even as the price of both commodities increased. I do not think that any useful predictive power exists for forecast horizons of less than a couple of years, so some long time period smoothing might help reveal underlying trends."