Dan Lippe
Petral Consulting Co.
Houston
In March-April 2020, governments around the world issued recommendations and mandates to limit social interaction to slow the spread of coronavirus (COVID-19). The world community became familiar with new terminology including self-isolation, social distancing, mask requirements, and remote operations. Governments also imposed nearly universal bans on cruise and international air travel. The shift to working remotely also caused a collapse in domestic air travel. The impact of these recommendations and mandates resulted in sharp declines in economic activity for a few segments of the population but not all.
For midstream companies, the global economic collapse resulted in reduced demand for gasoline, jet fuel, and diesel fuel. The domestic refining industry responded to the change in overall demand for refined products by reducing operating rates (especially in North America and Europe) and adjusting the mix of refined products. As refineries reduced operating rates, global demand for crude oil declined.
In the first few months of the global economic collapse, consulting firms and analysts within government agencies such as the US Energy Information Administration (EIA), International Energy Agency (IEA), and the Organization for Petroleum Exporting Countries (OPEC) were effectively flying blind.
Reliable domestic data from EIA—the most dependable and reliable data source for US petroleum products—routinely has a 2-month lag (January 2020 data was available on April 1, with June data published on Sept. 1). Data from the IEA and Joint Organisations Data Initiative (JODI) have similar lags but are less reliable because participation by many countries is voluntary. OPEC’s Monthly Oil Market Report (MOMR) is the most reliable source of data for global oil production but does not routinely track refinery crude runs, the all-important statistic for crude oil demand. The combination of OPEC’s MOMR report on global oil production and JODI’s monthly statistics for refinery crude runs provide the best insight in variations in global crude oil supply-demand.
Acknowledging the importance of trends in international demand for refined products and crude oil production, this article series focuses on the US domestic midstream industry. A special article focusing on the global market will follow separately.
US refinery crude demand
Since 2016, variations in crude oil and associated gas production have been the dominant variables that determine US gas plant NGL production volumes, particularly in the key oil producing shale basins (Eagle Ford, Permian, and Bakken). Forward views on crude oil and associated gas production will determine gas-plant NGL supply for the immediate future and its resulting impact on the midstream distribution system in eastern US markets.
Within the US, the impact of decline and recovery in demand for gasoline, jet fuel, and distillate fuel oil varied regionally. Refinery crude runs have natural seasonal variability. Total US crude runs almost always decline during the transitional months between winter and summer. In 2015-18, US refinery crude runs in winter months (October-March) were 300,000-400,000 b/d less than in preceding summer months (April-September). In winter 2019, refinery crude runs were 818,000 b/d less than in the preceding summer. Before governments officially suspended international air travel in March-April 2020, domestic demand for jet fuel was already declining. Domestic demand for jet fuel in first-quarter 2020 was 194,000 b/d (11%) less than fourth-quarter 2019 and was a leading indicator of what would occur in second-quarter 2020.
As local, state, and federal governments began recommending and mandating closures of churches, restaurants, and bars, demand for gasoline and distillate fuel oil also declined. Domestic refining companies began adjusting crude runs and product slates in ways that were almost unimaginable before second-quarter 2020. While the details are fascinating and important, the trend in refinery crude runs in first-half 2020 is representative of the overall impact of voluntary and mandatory closures on refined products demand in all US markets.
Table 1 shows US refinery crude runs, by region, first-quarter 2019 through second-quarter 2020.
Regional crude production
Many observers, journalists, and analysts review and comment on Baker Hughes’ weekly rig count report. While weekly variations usually result in limited commentary in general and industry press, long-term trends are always useful. In 2012-19, independent oil producers dominated oil-directed drilling activity in the major oil-prone shale plays. Independent oil companies rely on internal cash flow and external funding from hedge funds, private equity investors, and bank loans. In 2012-18, fund managers strongly encouraged independent producers to emphasize production growth. Production growth and growth in net revenue (free cash flow) went hand-in-hand when West Texas Intermediate (WTI)-Cushing oil prices were $60-80/bbl. Sustained recovery in the oil-directed rig count and rising crude oil production support the premise that the financial community viewed the drop in crude oil prices in 2015-16 as an anomaly and continued emphasizing production growth as the primary objective.
With sustained support from the financial community, independent US crude oil producers continued operating daily with an intense focus on production growth until late 2018 or early 2019. Oil-directed rig counts in the major shale plays began declining in early 2019, with rig counts for Eagle Ford and the Permian 16-18% lower in December 2019 than January 2019, and rig counts for Woodford shale in Oklahoma down 61%. During the same period, the rig count in Williston basin fell only 8.4%.
Rig counts in all oil-prone shale basins were in free fall in second-quarter 2020, but some basins fared better than others. Rig counts in Eagle Ford, Niobrara, Williston, and Woodford in August 2020 were 80% less than in fourth-quarter 2019, while the Permian basin rig count was only 70% lower. In the previous oil price crash (October 2014), rig counts began falling in fourth-quarter 2014, but the full impact of falling prices was delayed until second-quarter 2015, with rig counts continuing to fall through third-quarter 2016.
Fig. 1 shows rig counts for the four major US basin from January 2019-August 2020.
Production in the Permian, Eagle Ford, Bakken, and Niobrara was 6.4-6.7 million b/d in first-quarter 2019 and had increased to 7.3-7.5 million b/d by November 2019-February 2020, with production in February 2020 1.03 million b/d more than in January 2019 (Fig. 2). During this same period, US oil production increased 970,000 b/d, with Permian production rising 835,000 b/d to account for 81% of the increase in the four major shale plays and 86% of growth in total US production.
In March 2020, as refiners began cutting crude runs and before oil prices collapsed, producers in these four basins reduced production by 460,000 b/d. Bakken production fell 244,000 b/d (16.8%) while Permian production was 94,000 b/d (2.3%) lower and Eagle Ford production down 79,000 b/d (6.3%). US East Coast (USEC) refinery crude runs continued to decline in April-May 2020 but began recovering in June.
Fig. 3 compares US shale basin production with USEC refinery crude runs.
Regional variations in crude production are important as the drivers for associated gas production and gas-plant throughput rates. These variations are also important because gas quality (NGL content/MMcf) varies. Due to its volume of oil production, however, the Permian basin will be the primary driver of US output for the next 3-5 years.
NGL raw-mix, refinery propane production
At publication of the previous midstream article in this series, NGL production statistics were only available for January 2020 (OGJ, June 1, p. 39). Estimates of propane+ production (NGL raw mix excluding ethane) were based on EIA weekly crude oil production statistics and reasonable estimates for oil production in New Mexico, North Dakota, Oklahoma, and Texas. At that time, EIA’s weekly statistics showed US oil production was steady at 13 million b/d in March but fell to 12.2 million b/d in mid-April. Weekly statistics were a reasonable first estimate. As the oil price collapse forced producers into survival mode, the weekly oil production statistics became noticeably less reliable as the basis for estimates of propane+ supply. EIA’s weekly statistics showed oil production was 11.6 million b/d in May and 10.9 million b/d in June. By release of finalized statistics in EIA’s Petroleum Supply Monthly report on Aug. 31, 2020, data showed oil production was 9.4 million b/d in May and 10.4 million b/d in June.
EIA has now released monthly statistics for first-half 2020. NGL production from all US gas plants was 5.2 million b/d in first-quarter 2020 before slipping to 4.7 million b/d in May. As crude oil production in New Mexico, North Dakota, Oklahoma, and Texas recovered, gas-plant NGL production rebounded to 5.2 million b/d and was 4.9 million b/d for second-quarter 2020 (Fig. 4). Propane+ production (strongly correlated with oil production for all basins except Marcellus-Utica) was 3.2 million b/d in first-quarter 2020 before declining to 3.0 million b/d in the second quarter (Table 2).
Propane+ ratios vs. US oil production in May-June 2020 were 15% higher than in January-April. Production trends for propane+ and crude oil for Texas, New Mexico, North Dakota, and Colorado-Wyoming showed similar changes in production ratios. The increases in production ratios for Texas, New Mexico, and North Dakota can be explained by the shift in associated gas production from flaring at the wellhead to delivery to local gas processing plants. Flared gas volumes, however, were insignificant in Colorado and Wyoming in the early months of 2020.
The outlook for gas-plant NGL production for the next 18-24 months is more uncertain than usual for several reasons. Most importantly, unless oil production in these four states recovers to first-quarter 2020 volumes, reduced oil production will eventually translate into reduced volumes of associated gas deliveries to gas processing plants. In the meantime, Petral Consulting Co. has a moderately optimistic view of continual recovery in US oil production and gas-plant propane+ supply. Based on 12.25-12.50 million b/d of US oil production, propane+ supply will average 2.8-3.2 million b/d in second-half 2021. When ethane supply volumes are included, second-half 2021 US NGL production will average 4.8-5.2 million b/d.
Impacts on retail propane
Although US NGL production was more resilient than generally anticipated, propane supply from US refineries declined in all regions. In first-quarter 2020, refinery propane supply was 245,000 b/d, up 12,000 b/d from 2019. As refinery operations slowed amid reduced demand, operating rates for FCC units, reformers, hydrotreaters, and cokers also declined, reducing by-product supply of light hydrocarbons by 16-20%. Propane supply in second-quarter 2020 was 50,000 b/d lower from the first quarter and 60,000 b/d (20%) less than in 2019. Propane supply from USEC refineries during second-quarter 2020 was down 8,700 b/d from overall 2019 supply and 42% from second-quarter 2019.
Reacting to real and perceived supply shortages due to COVID-19, the propane wholesale marketing community scrambled to lock in supply for the coming winter heating season. While in a typical year propane inventory in all US primary storage sites increases 10-15 million bbl in April-May, in April-May 2020 inventory increased only 6.7 million bbl. During the same period, propane supply from gas plants and refineries was only 500,000 bbl less than in 2019, but exports rose by 3.2 million bbl. Propane imports from Canada, however, fell by 3.4 million bbl.
Propane exports in June 2020 were down from 2019, prompting inventory to increase to 9.5 million bbl in June and 10-11 million bbl in July. While the build in seasonal inventory in US Midcontinent and USEC storage during second-quarter 2020 equaled that of 2019, build for inventory in US Gulf Coast (USGC) storage was down 3.7 million bbl. EIA’s weekly statistics for July-August 2020 showed above average inventory build rates in USGC storage, offsetting weak build rates in the second quarter. Inventory in US storage sites is now on track to reach a seasonal peak of 100-105 million bbl on Nov. 1, 2020.
Propane inventory in Canadian storage increased almost 14 million bbl March-August 2020, and as of Sept. 1, 2020, was up 8 million bbl from 2019 and 3-5 million bbl higher than any previous year’s seasonal peak. In April 2020, Canadian National Railway Co. (CN Rail) reported it had laid off 16% of its staff, parked 500 locomotives, and shut down four train yards due to COVID-19 market impacts.1 Since Canadian propane marketers heavily depend on rail transportation, the deterioration in CN Rail’s service contributed to a drop in propane exports to US markets and abnormal inventory increases. A large fraction of propane inventory in Canadian storage very likely will be unavailable to ship to US markets in winter 2020.
Ethane recovery, rejection
An analysis of ethane production and rejection completes the current review of US NGL production.
From second-half 2019 through first-quarter 2020, when all gas plants operated at full recovery, maximum ethane supply exceeded demand by 0.95-1.2 million b/d (Table 3, Fig. 5).
Determined by three variables (gas-plant shrinkage costs, pipeline transportation costs, and fractionation fees), ethane recovery costs vary widely between two groups of NGL producers: independent gas processors and vertically integrated midstream operators. Independent gas processing companies must pay full pipeline tariffs and full contract rates for fractionation. Ethane producers belonging to the latter category of vertically integrated midstream companies—which own raw-mix pipelines and merchant fractionation—can choose to use variable operating costs for pipelines and fractionation when evaluating ethane recovery costs for their gas processing plants.
For the purposes of pro forma economic analysis, raw-mix pipeline tariffs of 5.5-6.5¢/gal apply to all Permian basin gas plants, while a tariff of 10¢/gal applies to all gas plants in Colorado and Wyoming. Fractionation fees are 5-6¢/gal. All ethane production from gas plants in New Mexico, Texas, Colorado, Wyoming, and Kansas-Oklahoma is transported to Conway, Kan. or Mont Belvieu, Tex. via raw-mix pipelines. Whereas gas plants in Colorado and Wyoming have the highest full-recovery cost basis, Permian gas plants have the lowest cost basis. Recovery costs in the Rocky Mountains are generally 3-4¢/gal higher than in the Permian.
Consistent with higher recovery costs and weaker margins, ethane rejection in the Rocky Mountains was 63-68% of estimated full-recovery volumes in second-half 2019 and first-quarter 2020 before falling to 54% of full recovery during second-quarter 2020. Ethane rejection in Texas-New Mexico was 22-24% of full recovery in second half 2019 and 20% in first-half 2020. Ethane rejection in low and intermediate-cost basins (Texas-New Mexico, Kansas-Oklahoma, Colorado-Wyoming) was 587,000 b/d in second-half 2019 and 533,000 b/d in first-half 2020.
Based on gradual recovery in crude oil production in second-half 2020 through yearend 2021, Petral Consulting estimates full recovery will average 2.6-2.9 million b/d, while ethane rejection will average 300,000-350,000 b/d in first-half 2021.
NGL market overview
Three markets account for more than 90% of US NGL demand:
- Petrochemical feedstock.
- Gasoline blending.
- International exports.
Export to a variety of international markets is the largest end-use for US propane and butane, and chemical feedstock markets (ethylene, propane dehydrogenation, specialty chemicals based on isobutane feed) remain the largest volume market for total US NGL supply (ethane, LPG, and natural gasoline). All five NGL components are used as feedstocks in petrochemical production. Normal butane, isobutane, and natural gasoline are used in gasoline blending (Fig. 6).
Beginning in second-quarter 2019, NGL exports set record highs for four consecutive quarters. NGL exports were 1.93 million b/d in second-half 2019 and 2.17 million b/d in first-quarter 2020 before dipping to 1.97 million b/d in second-quarter 2020. LPG exports (propane, normal butane, isobutane) were 1.4-1.5 million b/d in second-half 2019 and reached a record high 1.6 million b/d in first-quarter 2020 before slipping to 1.5 million b/d in second-quarter 2020. During second-half 2019 and first-half 2020, exports accounted for 38-43% of US gas plant NGL production.
Unsurprisingly, propane remained the most important NGL product export in first-half 2020, increasing 34.3 million bbl (18.4%) from the first 6 months of 2019. Butane exports during the same period were 15.4 million bbl (33.2%) more than in first-half 2019. Regionally, propane exports from USGC terminals were 101.7 million bbl in first-quarter 2020, up 26.1 million bbl from first-quarter 2019. In second-quarter 2020, exports slipped to 85.6 million bbl, down 1.05 million bbl from second-quarter 2019. Propane exports from USEC terminals were 10.3 million bbl in first-quarter 2020 (4 million bbl higher than first-quarter 2019) and 13.8 million bbl in the second quarter (down 120,000 bbl from second-quarter 2019). For winter 2020, international LPG buyers must consider two important questions: Will crude oil production curtailment continue to restrict Middle East LPG supply and exports? And, with US propane inventory at record-high volumes, will US exports increase enough to offset reduced shipments from Middle East producers?
The overview of domestic end uses focuses on chemical feedstock markets and refinery purchases (butanes, natural gasoline) for gasoline blending.
Petrochemical feedstock demand
Petral Consulting determines monthly ethylene production and feedstock demand by conducting an independent monthly survey of plant operating rates and feed slates to track consumption of purity ethane, ethane-propane mix, ethane, and propane contained in refinery offgas, purity propane (HD5), normal butane, natural gasoline, refinery-sourced naphtha, and gas oil.
US ethylene capacity reached 86 billion lb/year in first-quarter 2020. Nameplate capacity at yearend 2019 was 23 billion lb more than yearend 2015. Of this increase, most new plants were designed to run with 100%-purity ethane feed. During this period, producers in the USGC also retrofitted several multifeed plants to enhance purity ethane use.
Operable ethylene production capacity remained at that the 86-billion lb/year level until late August, when Hurricane Laura caused extensive damage to installations at Lake Charles, La. By late-September 2020, none of the ethylene producers were reporting extensive damage to ethylene plants or derivative units, but all plants and refineries remained shuttered until widespread damage to the regional electric power grid was fully repaired and power finally restored. As of Aug. 26, ethylene capacity in Lake Charles stood at 10.1 billion lb/year, accounting for 11.8% of total US capacity and 41% of capacity in Louisiana.
Feedstock demand for NGL feeds (ethane, propane, normal butane, natural gasoline) averaged 1.85-2.14 million b/d in first-half 2020, 2.08 million b/d in the first quarter and 1.97 million b/d in the second quarter. Demand for NGL feeds reached record-high volumes in first-half 2020, providing quantitative evidence that the global economic shut down had a limited impact on US chemical production.
Ethane demand during first-quarter 2020 was 1.74 million b/d and 1.62 million b/d in the second quarter. Ethane accounted for 78% of demand for all NGL feeds in the first quarter and 75.6% in the second quarter. Producers used 343,000 b/d of LPG feeds in first-quarter 2020 and 355,000 b/d in the second quarter, accounting for 15.5% and 16.6% of quarterly total NGL feed demand, respectively (Table 4).
Petral Consulting expects total demand for all NGL feeds will average 2.0-2.1 million b/d. If producers were able to run at 90-92% of nameplate capacity industry-wide, demand for fresh feed would increase to 2.5-2.6 million b/d. The difference between production rates of 170-190 million lb/day and 90-92% of nameplate capacity is 25-40 million lb/day (equivalent to 3-4.5 worldscale ethylene plants, each with a 3.3-billion lb/year nameplate capacity).
Refinery butane demand
Gasoline production is the number one factor in determining refinery demand for butanes and natural gasoline. In second-quarter 2020, gasoline production fell 1.79 million b/d (19.2%) from the first quarter and was 2.6 million b/d (26%) lower from the same period in 2019. In first-quarter 2020, gasoline yields accounted for 58.9% of crude runs before falling to 56.1% during the second quarter, when they were also 1.5-3.0% lower than in the previous 3 quarters.
The RVP-blending season of winter 2019 (October 2019-February 2020) was winding down as COVID-19 was emerging as the first pandemic since the Spanish flu of 1918-20. During April-September 2020, US refiners in all regional markets—as well as refiners across the globe—operated capacity at reduced rates of production. Since RVP-blending demand for gas-plant normal butane was already at seasonally minimum volumes by the end April 2020, gas-plant normal butane demand fell to 55,000 b/d during the second quarter from 228,000 b/d in the first quarter. Demand for gas-plant normal butane in first-half 2020, however, was up 0.5% from first-half 2019.
In sharp contrast, refinery demand for isobutane was 204,000 b/d in first-quarter 2020, down 6,700 b/d (3.2%) from first-quarter 2019. As refinery operating rates fell to their lows in April-May, isobutane demand averaged 109,000 b/d in April before recovering to 133,000 b/d in May and 170,000 b/d in June. Demand in the second quarter was 169,000 b/d, or 42,000 b/d (19.8%) less than the same period in 2019.
Refinery demand for natural gasoline fell more sharply from first-quarter 2020 to the second quarter than demand for butanes or gasoline production. First-quarter 2020 natural gasoline demand was 134,000 b/d, up 13,000 b/d (10.8%) from first-quarter 2019. As gasoline production fell, (particularly from USGC refineries), demand for natural gasoline dropped to 77,000 b/d, down 81,000 b/d (51.3%) vs. first-quarter 2019. The year-over-year decline in USGC refinery demand was 59,000 b/d, accounting for 74% of the overall decline. During second-quarter 2020, demand reached a low of 69,000 b/d in April before recovering to 80,000-81,000 b/d in May and June.
Near-term outlook
Petral Consulting’s paramount objective for each article in its series is to reduce the fog of uncertainty that influences decisions to capitalize on some opportunities and defer decisions on others. No matter how historians debate the circumstances of 2020, we learned some important lessons. Refinery operations managers, relying on software tools (e.g., linear optimization models for overall refinery operations and non-linear optimization models for gasoline blending), adjusted their product mixes to minimize impacts of the collapse in jet fuel demand.
As jet fuel demand declined, refinery managers reduced production rates by shifting the jet fuel fraction into the distillate fuel oil pool. Jet fuel yields were 9-11% in first-quarter 2020 before falling to 3.9% in May for a second-quarter average of 4.7%. At constant jet fuel yields, jet fuel production would have been 1.35 million b/d (735,000 b/d more than actual production). Distillate fuel oil yields were 30.8% in first-quarter 2020 and jumped to 36.8% in the second quarter. At constant yields, distillate fuel oil production would have been 4.05 million b/d (787,000 less than actual production). The drop in jet fuel production and the increase in distillate fuel production were offsetting.
Two questions will influence short-term marketing plans and longer-term strategic and capital budget plans for the next 5-10 years.
First, how long will Russia and Saudi Arabia remain committed to rebuilding their respective shares of the global market for crude oil exports? For almost 50 years, the key members of OPEC had a common objective: to limit crude oil production and defend a hard floor for benchmark crude prices, as higher prices are always better. In 2014, in the face of accelerating growth in US crude oil production, Saudi Arabia declined to lead OPEC to curtail crude oil production and sustain benchmark prices at $90-100/bbl. The objective, stated or unstated, was to halt growth in US crude oil production.
Second, how often will Saudi Arabia open the taps and collapse crude oil prices to achieve the dual objectives of maintaining discipline and high levels of compliance with OPEC production quota agreements and limiting growth in oil production from major US shale plays?
Those who manage day-to-day production, distribution, and marketing functions may be excused for concluding that Saudi Arabia’s leadership had reverted to the tried-and-true objectives and tactics of the previous 45 years. Others with longer-term responsibilities, however, have no such excuses, despite the fact that very few of us could have anticipated a pandemic that could push the global economy into its worst—and nearly instantaneous—collapse in any of our memories.
For the next 5-10 years, many will now constantly consider the prospect that OPEC’s new objectives are to make crude oil supply plentiful, manage the global supply-demand balance to create a hard cap on prices, and limit the likelihood that prices will increase to levels that erode market share for the world’s major exporting countries.
References
- Canadian National Railway Co., “CN Delivers Solid First Quarter Results Despite Network Disruptions,” Apr. 27, 2020, https://www.cn.ca/en/news/2020/04/cn-delivers-solid-first-quarter-results-despite-network-disrupti/
The author
Daniel L. Lippe ([email protected]) is president of Petral Consulting Co., which he founded in 1988. He has expertise in economic analysis of a broad spectrum of petroleum products including crude oil and refined products, natural gas, natural gas liquids, other ethylene feedstocks, and primary petrochemicals.
Lippe began his professional career in 1974 with Diamond Shamrock Chemical Co., moved into professional consulting in 1979, and has served petroleum, midstream, and petrochemical industry clients since. He holds a BS (1974) in chemical engineering from Texas A&M University and an MBA (1981) from Houston Baptist University. He is an active member of the Gas Processors Suppliers Association.
