Breakdown, It's All-Right. Impact of New EIA Natural Gas Storage Regions on Storage Predictions

Analyst estimates for this week’s Energy Information Administration (EIA) Weekly Natural Gas Storage Report before its release were rallying around an expectation of a 95-Bcf injection, according to the Wall Street Journal’s survey of storage analysts. The actual number reported by EIA yesterday (July 16, 2015) was a 99-Bcf injection, more or less in line with analyst expectations. But predictions may get a bit harder later this year. The EIA is preparing to redraw its US natural gas storage map and begin reporting inventory data in new regions later this year (2015). In August, prior to the launch of the revamped report, it will release a file with historical data for each of the new regions. The historical data will for the first time allow modelers to run their regressions and gather statistical information by which to rebuild their storage models designed to foretell the weekly EIA storage number. In the meantime, we did our own unscientific analysis of the regional breakdown and how it will change transparency in gas storage activity. Today, we examine storage capacities in the old versus new regions and potential impact on analyst visibility.

Recap

In Part 1 of Breakdown, we reviewed the new regional structure that EIA plans to launch sometime this fall following a period of testing. The EIA is planning to begin reporting natural gas storage inventory data for five US regions fashioned out of the three macro regions used currently. The data will be reported weekly in EIA’s Weekly Natural Gas Storage Report, which is, as we’ve noted in the past, the most closely watched piece of fundamental data in the natural gas market. We discussed the importance of this weekly report to gas market trading activity and why the regional changes are a big deal:  the new regional definitions will bring increased transparency and granularity to the gas storage data. But just as important to price and trading activity is the cottage industry of storage analysts looking to peg what the weekly storage data will report (see Trials and Tribulations of Predicting the EIA Natural Gas Storage Number). Thus, on a more practical level, we expect that the new regions will cause a temporary shake-up for those who build models to predict the outcome of that weekly storage data as they recalibrate their calculations around these new regions.

As we noted last time, the improvements will provide more precise inputs for the storage models. But assumptions and calculations behind the models will change. Exactly how the changes impact models and prediction accuracy will become clearer in the weeks and months following the launch of the new report. But we do know that one of the biggest factors in anticipating the weekly storage number is the degree of visibility modelers have into storage activity within each region. Analysts don’t see all the field-level data that EIA does but rather are restricted to publicly reported data, which is only a relatively small subset of storage operators in the EIA sample. (Gas storage operators are not required to publicly report storage inventory data.) The trick for analysts is to model the visible data to predict the storage numbers analysts can’t see in hopes of using that information to steal a march on the market. To that end, analysts use a combination of data, including production volumes, pipeline flows and weather, as well as historical correlations to fill the gaps and overcome the visibility issue. Thus, direct visibility into storage activity is just one factor affecting model accuracy. But, generally speaking, the lower the proportion of visible data compared to overall storage capacity – the greater the risk of estimation error. Because the new regional breakdown is going to change the analysts’ visibility by region it is therefore important to understand which regions will become more or less transparent after the changes.