Forecasting NatGas Prices With Fundamental Supply-Demand And Atmospheric Temperature Data

by: Robert P. Balan

Basically the fundamentals of NatGas is, like in any resource, the variance of current supply versus current demand. Add the element that imparts high-frequency price change: atmospheric temperature.

Like in oil, supply is not constrained (assuming no cartel machinations) but there's a short lag between production and availability if stored gas is not sufficient to meet immediate demand.

Tim Kiser and I came up with two approaches in trying to anticipate or "model" the likely future behaviour of NatGas prices.

If we do see a mover lower to 4.20 to complete an Elliott Wave triangle, we will pull the trigger for long NatGas.

We look for 6.00 optimal on the upside, but if the move becomes hyperbolic, we hope to tag along even higher.

Here is the prevailing meme in NatGas, provided by response to this comment from Seeking Alpha reader Open Mind:

Price dynamics of NG are very different than oil and in some sense completely removed from fundamentals. NG historically has been very cheap as a byproduct of oil. Later, with growing usage of NG and falling production of oil it has become a commodity produced for its own merits and priced accordingly.However, since shell oil revolution it was downgraded back to a byproduct. Oil producers have to sell it to avoid flaring which is limited by environmental concerns (quite reasonable in this case).

If they cannot sell NG and are not allowed to flair it they may be forced to shut in production of oil! In this case they may be willing to give it away for free, heck, even pay for it taking it away (negative price).

This discourages production and depresses storage - why invest in storing a fuel which may be impossible to sell profitably.

At the same time usage goes up very rapidly for electricity generation, home heating, chemical industry and LNG export.

As a result storage is rather tight and even moderate weather fluctuations cause elevated price swings. Currently high price of NG is purely a product of sudden fears and hedging related to potential shortage of NG in case of a cold and long winter. Nothing else - besides weather the demand is growing steadily and production is still abundant and cheap.

If the weather remains cold NG prices may easily double from current level, if it gets warm, it will go back below half of what it is. Sooner or later it will go down.

Open Mind described some of the contradictory imperatives in NatGas production, storage, consumption and subsequent pricing, and generally described as "conventional wisdom". But in the final analysis, looking beyond accusations that the NatGas market is generally rigged, this market responds incredibly well, and quickly, to the economic Law of Supply and Demand." And we add to that -- to the vagaries of weather.

I was pointing that out to members of the PAM community who may be not be aware that there is a NatGas financial variable that can be traded profitably, if prevailing "conventional wisdom" is put on hold, and one sticks to the pure fundamentals. The tools described below were developed to meet the interest of the PAM community to the potential of NatGas as a trading or investment instrument. Indeed, this year's internal price structure may be somewhat unique as illustrated by the NatGas calendar strip (see chart below):

Off the bat, we see how the current price rise is exceeding all expectations built into the price strip by a wide margin. That is a recipe for extreme volatility in the near future. Here is one of the underlying reasons why we are seeing this at this time:

For the 2108-2019 winter season, consumption has been outpacing supply to the tune of 650 - 750 Bcf over the five-year average, and that trend is expected by the models to persist until February next year. What exacerbates the supply-demand balance is that the the NatGas working inventories this season are well below those of the past seven years (see chart below).

Even as supply constraints roil the NatGas market, colder atmospheric temperature made its appearance earlier than usual, requiring net inventory withdrawal that is basically the strongest over the past seven years (see chart below). This is causing a classic squeeze on supply availability.

This classic supply squeeze has provided a lot of lessons for us at PAM management in developing the tools that we will describe in the latter part of this article.

Basically, the fundamentals of NatGas are, like in any resource, mostly revolves around the variance of current supply versus current demand. Just like in oil, supply is not very constrained (I am assuming no cartel machinations) but there is a short lag between production and availability if gas in storage is not sufficient to meet IMMEDIATE demand. That said, any large deviation from historical norms with regards to the four fundamental NatGas metrics (production, injections, net storage and net withdrawals, see chart below) can cause quick, or even lasting, price dislocations.

Immediate change in demand can be caused by radical change in atmospheric temperatures to high or low relative to benchmarked average temperatures or Heating Degree Days (HDDs), which by definition, is not generally expected. Heating degree days are a measure of how cold the temperature was on a given day or over a period of 30 days. It also used as a measurement designed to quantify the demand for energy needed to heat a building. Generally, HDD over its average course in any given year is periodic (seasonal), but a continuous average over 30 days or less, can display high frequency variability (as in the 30-day HDD of the US Northeastern Seaboard, light blue line, chart below).

Very high relative temperatures require NatGas to be consumed to power air-conditioning. Very low temperatures require NatGas for heating systems in homes, offices, factories, etc.

Therefore it is clear that: (1) if NatGas in storage is not sufficient to meet immediate and near future demands during extremes in atmospheric temperature, NatGas prices will tend to sharply escalate, sometimes with the price hyperbolic in its behavior; (2) there is a very distinct seasonality in NatGas price behavior as extremes in temperatures generally occur during summer and winter peak periods.

Tim Kiser (my partner at PAM) and I came up with two approaches in trying to anticipate or "model" the likely future behaviour of NatGas prices by:

(1) looking at the effect of the variance between Net Withdrawals versus Net Storage. Obviously if the yearly change rate in withdrawals exceed that of the change rate of storage, it is indicative of escalating demand, and obviously positive for NatGas prices. See chart below.

An unsaid point is that some decisions have to be made during the injection periods (done before peaks in atmospheric temperatures, usually in February or March) and that basically determine the availability of NatGas for immediate usage. Obviously, withdrawals are concurrent, and depends on the perceived or actual usage of NatGas needed to meet concurrent requirements.

If change rates in withdrawals exceed the estimated usage (which determined the amount NatGas in storage), you will see hyperbolic price moves corresponding to the variance. And we can see that dynamic in the general models of the variances in NatGas supply and consumption, demand, and inventories (see below). These models clearly define the periods which NatGas prices can reasonably outperform and provide optics as to the likelihood of that happening by using additive mathematics (see chart below).

We describe these as general models because they provide the directional price bent, but do not provide the high frequency turns provided by the changes in atmospheric temperatures (which impacts consumption). We need those inflection points to "trade" NatGas prices, instead of merely investing from the start of the season to its conclusion. Other investors may disagree over the objectives, but that is what we are aiming for.

We were therefore looking for models that can provide some clues as to the inflection points of future Natgas prices. This is how went about it.

The change rate in the supply-demand equation is usually ahead of the actual variances in the nominal quantities involved in the process. We also reasoned out that HDD, as a metric of what is essentially seasonal (thereby, periodic) vector, autocorrelation would provide some generic forecasts of the seasonality signature, which would of course improve as more data come in.

That is how we modeled those tools. In both instances, we opted for simple methods which use the raw NatGas fundamental data, manipulated only by transformation to their first derivatives to "normalize" relationships which allow graphical observation and analysis. The other approach is:

(2) to try to get a bearing on what to expect from weather changes, and then track the behavior of NatGas demand as proxied by Net Withdrawals.

The weather is unpredictable, pure and simple. But we can use a proxy for weather changes, then juxtapose the changes in withdrawals to this proxy, and see how they correlate. As proxy for weather behavior, we use the continuous average of Heating Degree Days (HDD) in the US NorthEastern seaboard.Then we take the final step of correlating the changes in NatGas prices to the optimized joint-behavior of the NatGas net withdrawals-HDD combo.

This is also what we see in the chart above. Rising prices until early February then falling prices until April-May. Sideways to slightly prices during June-July, and then flat to lower in August. By September, the situation becomes interesting again from higher price point of view. The model absolutely describes the seasonality of NatGas prices, which adds to our conviction that this method holds promise.

We extended the Net Withdrawals data further by autocorrelation (this should works, as the data set have very distinct periodicity), to get a feeling for what to expect 12 months ahead.

We are therefore tracking this Elliott Wave Triangle (see chart below), and if we do see a mover lower to 4.20 to complete a triangle, we will pull the trigger for long NatGas. Of course this is not the only possibility - the rally could blow away the upper trendline next week and proceed significantly higher - in which case, we will get aboard earlier.

We stick to the direction provided by the models, but we need to time the trades by employing market structure in deciding when to get in and when to exit. That is better than flipping a coin.

We are, therefore, waiting for opportunities to go long NatGas at lower levels. We may even get lucky.

We look for 6.00 optimal, but if the move becomes hyperbolic, we hope to continue tagging along even higher.

Disclosure: I am/we are long EQUITIES, OIL, PRECIOUS METALS, DXY. 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.