Today’s report (week ended 28 Dec): EIA: -20 bcf, EA: -30 bcf
- Today’s print missed consensus expectations by an unusually wide margin. We were on the low end of estimates, calling for a small pull, and had already built in an industrial demand impact of more than 1.5 bcf/d due to the holiday. To align with today’s print, we increased our production number, decreased the gas heating intensity value in the res-com sector and widened our holiday impact by 0.3 bcf/d. While power demand fell off due to milder weather, coal generation did not fall commensurately. Coal units had ramped up over the past month and operators were incentivised to keep them running due to onerous power market penalties if they were unavailable during a demand spike.
Next Thursday’s report (week ending 4 Jan): EA preliminary: -71 bcf
- Holiday-related impacts will remain in the industrial sector. Our model calls for a 1 bcf/d w/w recovery in the power sector and a nearly 10% w/w increase in GWHDDs, driving a surge in res-com demand. Importantly, there is a wide differential between our supply-demand balances which are calling for a 71 bcf withdrawal versus our flow model, which is calling for a much more tempered pull in the low 60s bcf, especially after today’s reported outsize injection into South Central inventories.
New Year’s evolution
The start of the new year has coincided with a new trading range for Henry Hub. Early in December, we noted that if a warming trend were to start in January, it could lead to the market anticipating a de-risking of deliverability concerns (see Panorama: Ho, ho, ho and a bottle of rum, 13 December 2018). The market has clearly priced in the significant amount of demand lost since the third week of December on warmth and the expectation for another paltry withdrawal next week. The potential for a shift to cold weather near day 10 in yesterday evening’s and this morning’s forecast that has been put forth by some commercial forecasters has ignited a small upward shift in the curve. At this point, the major bullish card for the market is whether this colder-than-normal period ultimately manifests and, if so, for what duration and at what deviation from normal. We have seen some prognosticators say that current guidance could lead to a period of cold through mid-March significant enough to erase this dead of winter warmth. In our opinion, only forecasts for a severe and sustained cold spell will significantly increase gas prices given the substantial inventory cushion in our balances that the December warmth and the warm start to January (including the upcoming five-day forecast) have provided.
Indeed, that forecast would have to be cold enough to return some 350+ bcf of demand to our balances. Our end-December storage projection has shifted from 2.57 tcf at the publication of our North America Outlook: Weathering the warmth, 20 December 2018 to now sit at 2.68 tcf. For January, using the current 15-day forecast and 10-year average weather for the remainder of the month, another 250 bcf or so would be lost in demand versus our projections based on 10-year normal weather. That is a substantial shift in just two weeks, with the likelihood for the staggering y/y storage deficit that has characterised balances to be nearly erased by the end of January if that potential cold does not materialise. If the current short-term weather forecast is realised, our end-March inventory projection removes most scarcity concerns, with an estimate north of 1.55 tcf. The current price hovering in the $2.90s-3.00/mmbtu range appears justified given the present short-term forecast, though we are still very cautious about what the impact that potential mid-January cold could have on balances.
Price weakness has migrated back into the injection season contracts. Our view has been (and still is) that 2019 injection season prices will ultimately depend on end-March storage levels. Currently, our balances place end-October 2019 inventories at just above 3.8 tcf. This figure makes an allowance for LNG maintenance in both April and October, and it suggests an injection season that will not be wanting for gas supply.
Beyond weather, we believe the injection season will also be characterised by significant demand-side timing risk from both LNG projects and pipeline exports to Mexico. This week saw a flurry of activity for approvals and requests for permission from FERC on the LNG front. Chief among the projects facing timing risks is Cameron LNG T1. Sempra Energy has said the train will have a Q1 19 in-service date, but the fact that Cameron LNG only received approval to commission its flare on 3 January and has not yet even filed for feedgas introduction points to delays into Q2 19. Kinder Morgan’s 10 mini-trains at Elba Island likewise face headwinds regarding start-up. The first train was scheduled to take feedgas in Q1 19 but will likely face delays after returning its compressor to the manufacturer for corrosion repairs in late December. Corpus Christi T2, which was approved to introduce fuel gas on 3 January, should be on track to take feedgas in late Q1 19/early Q2 19.
The developer of the Valley Crossing pipeline, the US-side of the massive 2.6 bcf/d Sur de Texas-Tuxpan pipeline project in Mexico still has not yet completed checks to identify subsea pipeline elevation as of 10 December, potentially delaying its Q1 19 start.
|Fig 1: Projected EIA weekly storage change, bcf||Fig 2: Projected storage y/y change, bcf/d|
|Source: EIA, Energy Aspects||Source: EIA, Energy Aspects|