However, starting with next week’s report we expect the year-on-year surplus to move back out as a result of last September’s shut-ins related to hurricanes, Gustav and then Ike. For instance, a year ago post shut-in injections dropped to 58, 67 and 57 Bcf respectively for the reports September 5 to September 19. If injections over the next three reports come in at the interpolated five-year average, the year-on-year surplus will move back out to 518 Bcf or 17% through the week ended September 18th.
To date, a total of 1,672 Bcf of molecules have been injected into underground storage. The EIA expects this refill season to end with a record 3.8 Tcf of gas in the ground. That would smash the previous record by ?260 Bcf or 7.3%. To get there, injections from this point only have to average 80½% of the 5-year average. Should we get to 3.8 Tcf, then as we look ahead to the coming heating season, withdrawals this winter will have to average 128% of the 5-year average or 2.41 Tcf to pull storage back to typical end-of-season levels before 2010 refills begin.
To put that into perspective, the largest winter withdrawal of working natural gas from underground storage on record occurred during the 2002-03 winter, 2.39 Tcf. And, if you recall, that winter was bitterly cold for the eastern half of the country; heating degree days ran between 5 to 25% above normal (see chart in today’s issue of The Schork Report ). Moreover, U.S. mills and factories were running around 75% of capacity then. Today, the U.S. factory economy is running at around 69%.
Oh yeah, and let’s not forget about the EIA’s latest 914-survey. Domestic production of natural gas in June unexpectedly rose by 52 bps to 63.3 Bcf/d. Furthermore, the estimates for April and May were revised higher. That is interesting given that the purge in the Baker Hughes rig count bottomed in between June and July at an average of 682 gas rigs employed. These rigs averaged 1,529 for the corresponding timestep from a year ago. To sum it up, gas rig counts this June and July dropped by 55% compared to 2008, yet year-on-year production as of June increased by 67 bps.
As such, key cross-seasonal time-spreads – the Oct’09/Nov’09, the Mar’10/Apr’10 and the winter/summer strips (Nov’09 to Mar’10) vs. (Apr’10 to Oct’10) – have been virtual freefall over the last month. In other words, the contango along the NYMEX forward curve is in freefall. For example, back in May when we started talking about the Oct’09/Nov’09 spread, the October contract was trading around 86 cents on the dollar to the November. Last night the October closed at 69 cents on the dollar. Over the same timestep the discount on the winter 2009-strip, relative to next summer has weakened from around 96 cents on the dollar to 88 cents as of last night!
Stephen Schork is the Editor of, "The Schork Report" and has more than 17 years experience in physical commodity and derivatives trading, risk systems modeling and structured commodity finance.