How well natural gas futures are in predicting the future?

In the last several days of February there has been a steady rise in natural gas future price (March delivery of Henry Hub), while spot price continue to decline, causing the gap between the two indexes to widen. At the beginning of the month it was the other way around and natural gas spot price was higher than future price.


Natural gas price chart – February prices


Natural gas spot price future (Henry Hub) Contango Backwardation February

The graph above shows the change I referred to.

The current gap between future (short term – one month ahead) and spot price is called Contango, which is reasonable to have, since future price includes additional costs above spot price e.g. interest rate. There are additional reasons for Contango, which include, among other, a rise in uncertainty level in the commodity, rise in expected inflation, etc.

It also might indicate, however, that the market prices the future more than the present, i.e. natural gas spot price in March might increase compare to February. This could explain the recent rise in future over spot price.

When the situation is in reverse, i.e. spot is higher than future, it is called backwardation.

Do future prices predict the future in a way that when natural gas switches to Contango, this situation could mean that the market (traders) expects natural gas spot price to be traded in the near future at a higher rate than now?

In other words how well the market does in predicting the shifts in natural gas prices?

To check this matter I have gathered data from 2010 and 2011 on natural gas spot price and future (short term) price of Henry Hub. I used only data from last year because, as stated above, there are many additional factors affecting future prices (e.g. inflation, interest rates) the shorter term we take the less likely these variables will affect. The tradeoff is using very few observations which jeopardize the robustness of the test.

Natural gas spot and future prices

The graph below shows the changes in natural gas Henry Hub future (short term – one month) and Henry Hub spot price over the years 2008-2011. It shows, as expected, a strong correlation between the two. Also the correlation of the monthly percent changes was 91%.


Natural gas spot price & future (Henry Hub) 2008-2011

When looking, however, at the graph below, which includes spot prices on a daily basis over the last year and backwardation/contango (red line) changes, there are two observations come into play:

  1. First, it shows a lot of noise…
  2. Second, it also shows that during the months of January- February 2010, when natural gas spot price was high, the backwardation was also peaking (i.e. future price smaller than spot price) and spot prices began to fall afterwards; and during October- November 2010, when natural gas spot price was low, the Contango was high (i.e. future price higher than spot price) and the prices started to rise in the following months.


Natural gas spot price (Henry Hub) Contango Backwardation 2010-2011

These observations needed to be tested:

Here below is a table that summarizes the findings:

Natural gas spot price 2010-2011 Contango Backwardation CORREALTION

The table shows the average prices of natural gas spot price (column D), their percent change (column E), the median Contango/Backwardation (future-spot) of daily prices (column C), and their lag by one period (column B).

If the market is right (or most of the time) then we should see a strong correlation between C and E, i.e. the contango/backwardation could fairly predict the following month’s spot price change, e.g. if during October   the median of the difference between future and spot prices was above zero (Contango, i.e. the market expects a rise in prices), we should expect to see a rise in natural gas spot price in November.

Unfortunately, the tables shows that the correlation between the two figures is only 10.8%, i.e. there is little evidence to support the claim that the market may predict the future, or at least it does so very poorly.

What then can we take from the long exercise?

For one thing, this doesn’t put the kybosh on the matter, only puts it on hold until further evidence will present its self;

And finally, when considering the recent changes in the gap between natural gas spot and future prices, one shouldn’t automatically conclude that if future price rises, it means the market values the commodity correctly and we might expect a rise in spot price in the next several weeks; it might just be the fear factor traders incorporate into future prices.

For further reading (on this site):