High elasticity of demand during the injection season makes natural gas trading more price-dependent and less weather-dependent (see: "The Logic of This Injection Season: What You Need to Know"). Therefore, lower prices create buying opportunities, not selling opportunities, as technical analysis often teaches us. Paradoxical though it may sound, lower prices are actually good for natural gas bulls. But what does history show? Can previous injection seasons teach us anything or maybe shed some light on what should be the price of natural gas this year? Let's have a look.
"There are three kinds of lies: lies, damned lies, and statistics". - Mark Twain
Statistics are never perfect, but this is the only tool we have that enables us to collect, organize, and analyze data. As a general rule, when we talk about any data, we usually talk about the past. The past is always stable, always reliable, and often very well-researched. Unfortunately, in some respect, it is also useless. Traders do no trade the past, they trade the future. And, the future is unknown; it is never stable and never reliable. The most popular way to predict the future is to extrapolate it from a set of historical data. This is true for both fundamental and technical analysis. However, as you well know, past performance is not indicative of future results. In other words, there is no sure way to know the future, but there are tools that can give you a rough idea about what it may look like.
In this article, we will briefly show what the future may hold for natural gas prices. We have made the following assumptions:
1. The price of natural gas is a reflection of supply and demand. In other words, the balance between supply and demand determines the price.
2. Storage figures (stockpiles or inventories) are the single most important piece of fundamental data in natural gas market. They have the strongest impact on the price because they show the net result of the interaction between supply and demand.
As already mentioned, history will be our only guidance. We will assume that natural gas price has a strong linear dependence on historical observations.
End-of-Season Storage Regressions
Basically, we took two variables: historical end-of-season storage figures (independent variable) and the closest price of the natural gas prompt month contract (dependent variable). In the case of withdrawal season, (EOwS), the price is for the May contract. In case of injection season (EOiS), the price is for December contract. We then run linear regression to get coefficient of determination (R squared, R2) between two variables. R2 ranges from 0 to 1, with zero indicating poor predictability and one indicating high predictability. Below are the results.
Source: Bluegold Research
As you can see from the chart above, there is a clear negative correlation between natural gas inventories at the end of withdrawal season and the price of May contract. The more gas there in storage, the lower is the price. The relationship, of course, is not perfect. R2 is around 0.62. However, when we remove the year 2015 from the model (which seems like an outlier), the coefficient of determination improves to 0.84, which is a very good number.
Source: Bluegold Research
In the case of injection season, the strength of the relationship is much weaker (R2 is just 0.3511). The year 2014 and 2015 look like outliers, but when we remove them, R2 improves to just 0.63. Still, we have to work with the data we have, even though it is not faultless. By using R2 of 0.3511 and our latest end-of-injection season index of 3,625 bcf, we get December 2017 contract price of $4.026 per MMBtu, which implies 27% growth potential from the latest December contract price. When using R2 of 0.63, projected December contract price increases to $4.553 per MMBtu, which implies a 44% growth potential.
Please understand that this is a very simple linear forecast based on a single independent variable and without any adjustments. Still, it enables us to get a sense of what the price of natural gas would be, if it were perfectly correlated with history. Adding additional regressors into the model would help improve its quality and also strengthen the reliability of its forecast. For example, it seems reasonable to infer that adding a production variable would align the results in a more linear way, thereby removing outlier years (such as 2015 in EOwS and 2014 and 2015 in EOiS).
We have a lot more confidence in our end-of-storage index than we do in its ability to predict linear, perfectly correlated prices. Of course, we doubt that December contract can rally 27% (much less by 44%) from today, but rally it should. Because of the high volatility of natural market, we update our indices on a daily basis. Under the latest forecasts and assuming that market wants to see 3,800 bcf in storage by the end of October 2017, we estimate balancing price to be above $3.20 per MMbtu in July and August.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.