Why the IEA and EIA have their supply/demand forecast for 2018 wrong

The market has become decidedly bearish towards oil prices in the past six weeks as the US EIA has released several consecutive weekly production increases. In case you don’t know the EIA publishes weekly estimates of US oil production in its petroleum inventory report and forward month estimates in its monthly drilling productivity report. For example, the December drilling productivity report will estimate January 2018 production.

Based on recent bullish US production growth estimates, the IEA has revised its supply/demand estimates to show a surplus of 300K throughout 2018. This estimate is decidedly bearish and should be of concern to OPEC in its effort to re-balance the market. The IEA bases its estimate of surplus on an assumption that worldwide production will increase by 1.6 million bpd against demand growth of 1.3 million bpd. The IEA estimates that 1.1 million bpd of new production will come from the US.

I along with others, have felt for a while that the EIA production estimates are too optimistic.  The November Drilling Report estimates that Permian production will rise in December by 59,000 bpd to 2.7 million bpd, a rise of about 650,000 bpd from 2.05 million bpd at the beginning of 2017, or a rise of 32% year over year. The Bakken, Eagleford and Niobrara per the EIA, have seen an aggregate increase of about 125,000 bpd in 2017.

Those of us with experience in engineering or production accounting know that production numbers aren’t official until they are issued by AER or TRC.  Final production numbers are used for royalty, tax and payment of partners.  The TRC publishes preliminary production numbers in the second month with final numbers published one month later.  So, to wrap up the debate of EIA estimates versus actual production numbers, TRC numbers are confirmed three months after the fact while the drilling productivity reports on the forward month, resulting in a four-month gap. In a rock, scissors, paper debate, state numbers win, always, except for federal offshore lands.

Next, I did a little fact checking on my way to concluding that the EIA estimates are likely too high. First, I looked at TRC data of oil and condensate production for January 2016, January 2017 and July 2017 (before the hurricanes struck). The numbers are 103,026,307 bbls (3,323,429 bpd), 96,970,239 (3,128,072 bpd) and 93,883,761 (3,028,508 bpd) respectively. These numbers from Texas support the belief that US oil production has not grown as rapidly as thought.

Because the Permian has been the main driver of growth, I looked at Permian production from two main sources; namely the TRC and the webpages of three independent Permian focused oil and gas producers. As shown in “Texas Permian Basin Oil Production 2008 through September 2017” (Source: Railroad Commission of Texas Production Data Query System (PDQ) dated November 30, 2017), 2016 Permian production averaged 1,556,717 barrels per day and 1,632,930 bpd YTD 2017, representing an increase of 5% (6.5% extrapolated to December 2017).  I did not look at detailed condensate production as it is a small fraction of oil production, and the larger picture was illustrated in the historical data above.

Actual producer data obtained from investor presentations are as equally powerful as government data in reviewing real world Permian growth, as the data must conform with SCC rules. The first producer I looked at is based in primarily in New Mexico. Its 2017 guidance ranges between 61,000 and 64,000 boepd.  Its YTD 2017 and Q4-2017 oil production grow by 4,350 bpd year over year an increase of 18% greater than in 2016.

Next, I looked at the leading Permian operator in Texas.  As outlined in its November corporate presentation, its oil production grew from an average of 134,000 bopd in 2016 by 17.5% to 157,500 bopd, or an increase of 23,500 bopd in 2017.  Finally, another top ten producer has actually seen its oil production marginally decrease from 79,751 bopd in 2016 to 77,071 bopd YTD to September of 2017.

I do not have access to nor have a crystal ball as to how the EIA computes its weekly and monthly projections of production. I can only guess that their model consists of type well data, with imputed declines and assumptions as to monthly rig releases and completions.  Based on the above data, and comments by industry leaders like Harold Hamm, I feel that the EIA numbers are flawed to a degree. There are many reasons for this including: non-type well production (one size does not fit all across huge basins), delays in facility tie-ins, train wrecks (wells not completed); facility bottlenecks and lower than assumed oil production.

I think at the end of the day, 2017 Permian production will be seen to have grown in a range between 15% and 18%, not the 30+% estimated by the EIA.  Assuming January production of 2.2 million bpd exit rate Permian production should fall within a range of 2.53 and 2.6 Million bpd, or net growth of between 350 and 400K bpd.

So how does affect US oil production in 2018, assuming similar activity rates? I think we will see between 500K and 600k bpd growth from the shale basins and maybe 100K from the Gulf of Mexico.

Bottom-line, the EIA and the IEA are off by up to 500K bpd in their 2018 production estimates. Depending on whether the ‘other’ 600K of new supplies from the rest of the world shows up, we could be in a for a wild ride in OECD inventory reductions and price increases.

One more thing to now, one of the aforementioned producers noted Permian drilling and completion costs now range between $7.7 and $8.9 million depending on TD and TVD. This suggests that most of the post price crash service provider discounts have been reversed.