The counter-seasonal 7 Bcf injection into natural gas storage reported yesterday by the EIA came in towards the pessimistic end of the estimate range. Yet, the ostensible "surprise" did not seem to cause much of a price reaction. After dropping momentarily on the announcement, the prompt month natural gas price returned back to the level that prevailed at the beginning of the day.
Some readers may be wondering, why the "surprise" large injection did not matter?
While trying to interpret intraday natural gas price moves is hardly a productive way to spend one's time, I will dare to suggest that the report indeed did not matter this week. And rarely does in general.
I invite readers to review a sample of select intraday price charts focused on EIA releases at 10:30am EST on Thursdays that I included in the Appendix below. While each of those charts shows a spike in price volatility and trading volume during the minutes following the data release, only on few occasions (specifically, on Feb. 9 and Feb. 16, out of a total of eight cases) the charts appear to show detectable residual memory of the event in the price direction after several hours of trading. After moving around with a wide amplitude immediately before and after the announcement, the price appears to return to the pre-announcement trend line more often than not.
This observation - admittedly based on an ad hoc sample and my subjective interpretation - has a rational underpinning, in my view. What is often depicted as a "surprise" by journalists and bloggers, in reality is most likely immaterial and non-actionable news to market participants who really matter: the well-capitalized dedicated natural gas trading desks.
To begin with, storage estimates derived from pipeline flow models are generally characterized by relatively high precision. Large natural gas traders typically pay for models provided by one or more of the several major consultancies and often have elaborate analytical infrastructure of their own (which may exceed some of the external models in quality). Given that the third-party models are distributed widely among the subscribers, information leakage is inevitable. Estimates are propagated via personal communications and blogosphere plagiarism and become "common knowledge" for a fairly wide category of natural gas traders (including smaller traders who are not prepared to pay the subscription fees).
By contrast, "consensus" estimates quoted by journalists often represent averages of estimates based on a wide spectrum of public sources. Some of those sources (such as, for example, equity-focused analysts covering the Energy space on the sell side) are not dedicated to natural gas research and cannot compete in precision with the specialized consultancies. As a result, the "consensus estimate" typically represents a mechanical average across an eclectic group of models that vary widely by type and in quality.
In this context, while the actual data point released by the EIA on a Thursday may appear to be a surprise event relative to such "consensus" benchmarks, it may be less of a surprise when compared to a set of more elaborate pipeline flow models.
There is another important aspect that needs to be taken in consideration.
Every model produces a prediction with a range of estimation error. For example, from a statistical perspective, an estimate of a 9 Bcf injection into storage for last week can be indistinguishable from an estimate of a 5 Bcf injection, even if derived from a well-maintained pipeline flow model.
Furthermore, the report by the EIA is also an estimate (albeit based on a survey). Taking in consideration the typical range of EIA's reporting error, the headline storage level of 2,363 Bcf may be reflecting a "true" underlying storage level of, let's say for illustration, 2,360 Bcf or 2,366 Bcf. While the precision of the EIA's weekly survey data points is quite high, it is important to remember that the headline injection or draw figure represents a difference of two comparable estimates and, therefore, has a relatively wide error range as percentage of the reported value. In other words, from a statistical perspective, a 7 Bcf reported injection can be almost the same thing as a 4 Bcf reported injection or a 10 Bcf reported injection (all figures are illustrative).
Once these natural imprecision ranges are taken into account and superimposed, it becomes clear why a variance between the model estimate and the reported data point is unlikely to be an actionable data point for a rational trader, even assuming high confidence in the model (However, big and persistent discrepancies could be a reason to switch to another consultant).
The graphs below appear to support this reasoning.
EIA's weekly reports remain an important source or public, almost real-time data, and therefore are quite valuable. However, in my opinion, knee-jerk trading decisions based on the perceived "surprise" vs. a consensus estimate can be a precarious path leading to possible costly disappointments.
OIL ANALYTICS updates its view on natural gas fundamentals on a frequent basis.
Appendix: Natural Gas Price Reactions To Weekly Storage Reports
Disclaimer: Opinions expressed by the author in materials included in Zeits OIL ANALYTICS subscription service or posted on Seeking Alpha's public site are not an investment recommendation and are not meant to be relied upon in investment decisions. The author is not acting in an investment, tax, legal or any other advisory capacity. The author's opinions expressed herein address only select aspects of potential investment in securities of the companies mentioned or commodities and cannot be a substitute for comprehensive investment analysis. Any analysis presented herein is illustrative in nature, limited in scope, based on an incomplete set of information, and has limitations to its accuracy. The author recommends that potential and existing investors conduct thorough investment research of their own, including detailed review of the companies' SEC filings, and consult a qualified investment advisor. The information upon which this material is based was obtained from sources believed to be reliable, but has not been independently verified. Therefore, the author cannot guarantee its accuracy. Any opinions or estimates constitute the author's best judgment as of the date of publication, and are subject to change without notice. The author explicitly disclaims any liability that may arise from the use of this material.
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.