Much has been made lately about the price ratio of WTI (per bbl) to Henry Hub (per mcf of natural gas). This is understandable given that most people treat the the idea that this ratio must return to 8.5 as a kind of received knowledge. Unfortunately, there has been little actual analysis that I've seen here. It is a bit disturbing that when reference is made to this ratio the discussion tends to center around which component of the ratio is likely to adjust rather than whether the ratio will actually adjust in a predictable way that can be traded. In deference to the trader attention span, I will give you the answer up front: no. If you care to know how I reached this conclusion, read on.
Fundamentally, there is little reason oil and natural gas should trade against each other in the near term in a meaningful way. US infrastructure (transportation, industrial, commercial, residential) is not generally designed for easy switching between oil (or its distillates) and natural gas. Some ability to switch between gas and 6 oil exists for industrial users, but not enough to have a significant impact on the marginal price of either commodity. A wealth of data supports this assertion, much of which can be found here.
So where did the idea that the WTI/HH ratio should alway trend toward 8.5 come from? The chart below provides insight.
By golly, that red average line is exactly 8.5. I guess that settles it. Time to head for the pub. On the other hand, maybe we should think about this a little more. There are so many questions to ask, such as "how stable is this average over time?" Let's see. The rolling average should do nicely.
Hold the phone, Mable! It looks like this average is changing over time. I guess we are going to have to dig a bit deeper.
What are we really trying to determine? If one is to use this ratio as part of the basis of a natural gas trading strategy, then the central question seems to be "does the ratio of WTI/HH contain information of predictive value for future natural gas prices?" Time to dust off the statistics book and perform some regression analysis. Below I ask and answer three questions using EIA data and simple linear regression tools.
Does the current month average WTI/HH ratio have value in predicting next month's average NG price?
A simple linear regression using a 1 month lag of the WTI/HH ratio as the independent variable yields a cooefficient of determination of 0.012. In other words, no.
Does the current month average WTI/HH ratio have value in predicting the average NG price six months from now?
A simple linear regression using a 6 month lag of the WTI/HH ratio as the independent variable yields a coefficient of determination of 0.000026. In other words, hell no.
Does the recent direction of change in the WTI/HH ratio have value in predicting next month's average NG price?
A simple linear regression using the delta of the WTI/HH ratio from (t-1) to (t) as the independent variable yields a coefficient of determination of 0.0013. So again the answer is no.
There are a thousand other iterations that we could go through, some of which might even appear to have some predictive value. These are likely to be coincidental. Neither the fundamentals nor statistical analysis support the hypothesis that the WTI/HH ratio is useful in determining NG prices. So unless you want to base your trading strategy on the hope for reversion to an unstable mean, it would be wise to stay away from this ratio.
Disclosure: No physical or financial positions in NG