Today’s report (week ended 28 Sep): EIA: +98 bcf, EA: +87 bcf
- Today’s report once more was outside the range of consensus. To align with today’s balances, we had to lower our estimate for gas-fired power burn by 1.3 bcf/d. Given that today’s release was the last for the reporting month, it could be an issue of the EIA ‘truing up’ for end of the month storage totals, compensating for last week’s low number compared to consensus.
Next Thursday’s report (week ending 5 Oct): EA preliminary: +89 bcf
- Next week’s report should show lower injection activity on a 0.3 bcf/d w/w decline in production and power burns that step up nearly 1.0 bcf/d w/w.
EIA Monthly – July
Last week’s release of the EIA’s Natural Gas Annual and Natural Gas Monthly once more underscored the incredible ongoing strength in US gas power burn. July data indicate a 39.6 bcf/d level of burn, around 1 bcf/d higher than our initial (but still incredibly strong) 38.5 bcf/d reading based on our models. That strength has continued, thanks not only to weather but to competing fuels prices as well (see E-mail alert: Fuels rally pushes power prices higher with gas-fired plants in Appalachia, Southeast to benefit further into 2019, 1 October 2018). Our models indicate gas burn was 34.5 bcf/d in September, and for October that figure is nearly 30 bcf/d, based on current weather forecasts.
There were no other surprises in the data, but the true-up process from the release of the EIA Natural Gas Annual resulted in a sizeable upward revision to domestic production. The EIA confirmed another massive m/m increase in production, of 1.6 bcf/d in July. A figure of that scale had been indicated by our pipeline scrape data. Since then, a series of ensuing sequential gains has likely increased production by another 3 bcf/d or so through end-September.
For October, our monthly balances had called for another 1.2 bcf/d m/m increase, but recent pipeline flow readings suggest a lower sequential gain. With Atlantic Sunrise receiving FERC approval to enter service on 4 October, Appalachian uptick could send Lower 48 readings back onto their previous trajectory.
Canadian net trade came in somewhat higher than our expectations. In July and for the second straight month, the disparity between our estimate for 5.8 bcf/d in net Canadian imports and the EIA’s figure of just under 6.0 bcf/d could be pinned on differing estimates of flows on the Rover pipeline. The EIA’s official data show that US exports to Canada through the St. Clair border crossing—where Rover flows through the connected Vector Pipeline cross into Canada—increased by just 0.1 bcf/d m/m, to 0.8 bcf/d. Our flow data indicated a 0.3 bcf/d m/m increase in US exports through St. Clair, to 1.7 bcf/d. While we had anticipated these flows keeping net Canadian imports flat m/m, the EIA’s more tepid Rover number pushed its total for net Canadian trade up by 0.2 bcf/d m/m.
Industrial demand continued to exhibit strength, with the July reading, which should show little in the way of a weather effect, up by 0.9 bcf/d y/y. The three-month moving average, which we use to cancel ‘noise’ in data, is now pointing to 0.8 bcf/d y/y of growth for baseload gas uses. The EIA data had those volumes ratcheting up beginning in May, which may be underscoring a dryer stream of gas for West South Central industrial users given the run-up in ethane prices. Capacity additions are also helping to drive growth, in addition to any potential BTU content-related uptick, as we pointed to in our Insight: Winter outlook: deliverability vs availability, 24 September 2018. For the heating season, we are calling for 0.6-0.7 bcf/d in baseload (ex-heating) industrial demand growth.
EIA’s data for net Mexican pipeline trade confirmed our outlook for only a moderate uptick in July net exports as the El Encino-Topolobampo and Nueva Era pipelines entered service (see E-mail alert: EIA data confirms only 4.7 bcf/d net trade with Mexico in July, 1 October 2018). For the heating season, we are still calling for 0.6 bcf/d in growth on a slow ramp-up of infrastructure. Verification of August data, which would show a full month of service on both pipelines, could add upside risk to our estimate for October―we currently expect 4.8-4.9 bcf/d to cross the border then.
While less important for the balances, lease, plant and pipeline fuel also saw time series revisions. On a net basis, however, the total demand from the category was the same.
Another week has come to pass, and our weekly balances are once more notching down our estimate of end-October storage. Currently, our forecast stands at 3.23 tcf. Based on the 20-day forecast and normal weather assumptions beyond that window, our preliminary weekly balances are actually showing injections through the second week of November – with our balances for the week ending 16 November showing a marginal build, which still would bring peak inventories below 3.3 tcf before withdrawals begin. Henry Hub cash for today’s flow date at $3.26/mmbtu is clearly showing a premium for directing gas into storage at such low levels. For the November contract, that largest nearby open strike is at $3.25/mmbtu with some 16,000 lots. During yesterday’s rally, prices had a hard time breaching that resistance level and with today’s bearish miss and a string of 90-type bcf injection for the next two reports, it may prove to be a hard level to break. The biggest open strike below current trading is $3.00/mmbu with over 30,000 lots and should keep prices supported in the retreat post today’s release.