Today’s report (week ended 31 May): EIA net change: +119 bcf, EA: +109 bcf
- For a second week in a row, today’s print was well outside consensus estimates near 109-110 bcf. Our flow model, at 111 bcf, pointed to the risk of a higher number. We had previously built in a 1.4 bcf/d holiday impact for the industrial sector, which we have now increased to a 1.9 bcf/d drop. Like last week’s print, we find this week’s EIA injection hard to justify in our balances.
Next Thursday’s report (week ending 7 June): EA preliminary: +98 bcf
- A 0.7 bcf/d w/w decline in production, a 2 bcf/d increase in power gas burn and a return to business as usual after the Memorial Day holiday drive the fall in injection w/w.
The price is light
2019 now contains three of the top five injections of all time, according to EIA data. Our June balances are looking looser y/y based not only on the forecast (and to-date) lack of heat in June, but also because of some hiccups in LNG feedgas takes, showing how even a small degree of variability at multiple facilities can swing demand. On a weather-normalised basis this week, balances are running 2.1 bcf/d looser y/y and if the 15-day forecast is realised, they could run more than 3.1 bcf/d looser y/y by the week ending 14 June. Our models point to an end-June carryout near 2.39 tcf, 20-30 bcf higher than the estimate released in conjunction with last week’s Panorama.
Yesterday (5 June) posted the highest CDD count (both on a population-weighed and electricity-weighted basis) so far this summer and allowed power burn to reach a projected 34.3 bcf/d. There will not be another day that warm until 20 June, according to the latest 15-day forecast, and that lack of cooling demand will weigh on the ultimate burn that can be achieved this month.
We have adjusted our June power burn number downwards to 32 bcf/d given realised temperatures and a 15-day forecast pointing to June CDDs 11% below the 10-year normal (and assuming 10-year normal weather beyond the forecast period). That delta is worth a nearly 2 bcf/d loss in demand versus our forecast before June began. We estimate added power sector gas burn of around 1 bcf/d for each fall of $0.25/mmbtu in the cash price.
In the Midwest markets of MISO and SPP, where lower-cost coal units burning PRB are at risk with prices below $2.50/mmbtu, a lack of overall power demand—due to mild weather noted above—and growth in renewables has meant both coal-fired and gas-fired generation have declined y/y. That has left gas prices to do the work to help gas maintain share in a declining market for megawatts.
In PJM, which covers Chicago to Washington, gas-fired generation has increased recently due to the widening spread between low-cost efficient CCGTs and local coal-fired units, which run more expensive fuel (see Fig 4). Since the start of April, gas-fired generation in PJM has been up by nearly 3 GW, while coal generation is down by more than 4.3 GW. This ‘switching’ from coal to gas has been driven by new CCGTs, of which 12 GW were added since the start of 2018. These new plants averaged 5.3 GW or 0.9 bcf/d since the start of April, up by more than 2.1 GW y/y (nearly the entire growth in gas-fired generation). Because these new plants are more efficient though, the displacement of coal with gas has been less intensive in gas consumption terms.
As we outlined in our alert (see E-mail alert: Arb for US freedom gas to Europe for late summer delivery may close, 3 June 2019), even in the $2.30s/mmbtu the TTF-Henry Hub arb has largely closed for Jul-19 and Aug-19 contracts, which could keep some US gas locked in that would otherwise have been destined for export later this summer. However, any locked-in volumes are likely to be relatively small given many cargoes should already be hedged. In addition, we do not foresee an actual shutdown of a US train and instead envision facilities just running below their nameplate capacities.
Feedgas volumes into LNG facilities have dipped over this past weeked after the Jul-19 arb closed. However, that appears to be a coincidence. Cameron LNG volumes began to ramp down and are now at zero, with Kpler cargo-tracking data indicating a tanker is still a four-day journey away, suggesting that lower feedgas intake is driven by inventory management or it could just be part of the usual variability in take associated with the commissioning process. Feedgas volumes at Cheniere’s Corpus Christi fell as the Texas Commission on Environmental Quality reported that a power loss resulted in a plant shutdown and subsequent plant restart lasting from early 31 May to 2 June. Of note, we have seen some very minor volumes (0.01 bcf/d) heading into Elba.
In the short term, two consecutive bearish reports will obviously weigh on sentiment, as is evidenced by this morning’s price fall. Given how hard both prints are to justify within the context of our balances, the reports also raise questions as to whether next week’s EIA report could see a revision, or if there is a collective model miss that will lead to a recasting of models. Part of that miss could be from the industrial sector, where March EIA monthly data posted lower than our expectations as manufacturing activity nationally has slowed this spring, declining in three of the past four months based on Federal Reserve data amid flooding in the Midwest and several auto plant closures. Additionally, EIA volumes for Mexican trade in March underperformed scrape samples by some 0.2 bcf/d that could be translated into the current month on that baseline. Even those adjustments combined, however, would not account for the scale of the past two weekly misses. Instead, the scale of the miss might point to a potential recasting of production estimates, which pipeline flow models are not able to capture, and possibly power, though our forecasts have been extremely close to realised figures in both of these categories of late.
|Fig 1: Weekly EIA storage change, bcf||Fig 2: Weather-adjusted balances y/y, bcf/d|
|Source: EIA, Energy Aspects||Source: EIA, Energy Aspects|
|Fig 3: MISO & SPP gas share of coal & gas||Fig 4: PJM gas share of coal & gas|
|Source: MISO, SPP, Energy Aspects||Source: PJM, Energy Aspects|