Those along for the energy trade, or in the industry, have gotten rather excited over the past few months of US rig drops. Rightfully so, good news in the energy space has been slow recently, and prices haven’t been encouraging, especially considering the ‘super-cycle’ theme rehashed ad nauseam in 2022.
But, it doesn’t matter. The US rig count is wholly inconsequential to the oil thesis, and matters increasingly less as OPEC flexes their market-control muscles. Bulls, in the first place, were expecting no growth (even decline) out of the United States — and, I would posit, that, the general incentives that come with being a public company, would result in production being stickier than some may think — meaning, rig counts can fall, and it be of little important to the larger thesis — both bullish, and bearish narratives. There is this anchoring to US field data, as if it’s an accurate read on the rest of the world, I get it, both free, and high frequency, it’s easy to check in on the Baker Hughes rig count, but in an era where the US is no longer the driving oil market force, the significance of US data fades. The rig and frac spread count mattered a whole lot more in 2015-2019, where the US was responsible for the majority of global production growth, though today, incremental production is coming from everywhere but the United States it seems.
And, if we want to talk about the rig count, we have to talk about the (arguably) more important frac spread count — so let’s start with that.
(Explanation of frac fleet for those unfamiliar: the frac spread count is the number of frac fleets currently active. While the rig count is important, drilling a well is only the first sub-surface step before you can bring production online — before you can start producing an unconventional well (most of the new production in the US), you have to frac it. If you drill, but don’t frac a well, it’s a DUC (drilled, uncompleted). So, you can have an infinite number of rigs, but without frac fleets, they are useless).
Post-2014, the rig count to frac spread ratio has been almost exactly 2:1. Today, there are 260 frac spreads active, which would imply a target rig count of ~520 — we are at ~740 active rigs today according to Enverus. In all likelihood, we build DUCs, as we have drawn a good number over the past few years, frac spreads fall to around 250 crews, and companies go back to pushing efficiency gains in this not-low-but-not-high price environment. To win, you either need to push higher EUR, or lower DCET (the two parts of FD F&D costs, and thus your recycle/profitability ratio) — so, I would expect some revisiting of cost control measures (the inflation story is mostly played out, sand is down, casing is down, steel is down 50% from 2021 highs (though still elevated over 2019 prices) to improve cash margins on the recycle ratio numerator. That’s fine — it’s what should happen when we are uncertain about prices (as an aside, probably good to avoid operators who are putting emphasis on a 2H23 oil price recovery and have used that as an excuse not to improve their business in the meantime).
The slope of the correlation between new forward production added, is much steeper for frac spread count, compared to the rig count (on the left) — which makes sense, you have to complete the well before you bring it online — which is to say, that frac spreads are more important to pay attention to here if you are seriously tracking US production. There are ~270 active frac crews right now (~300 two months ago), which is flat over the last 2 years, and seasonally, in line with the 10 year average. When you start to move below 250 frac spreads is when you sweat — and 700 total/500 oil on the rig side. There is some truth, I believe, to the shale degradation story — but I do think it’s overstated just given the recency bias people tend to place on what they see and read every day, so add 10% to those numbers if you want to feel safe.
While rigs have been dropping, they have been dropping from secondary plays. Rigs in prolific, tight oil plays have continued to remain strong. Tight oil rigs in the 4 main basins have seen their share of total US onshore rigs increase so far this year, as producers step away from drilling lower quality acreage.
Not as frequently discussed, is drilling efficiency — both new well production per rig, amd number of wells completed per rig, per month. 2016 through 2019, the average US rig was drilling ~1.6 wells per month, while today, that is closer to 1.8 wells per month, an almost 15% improvement. New well production per rig is also higher, thanks to longer wells and modern designs — but new well production per rig has softened over the past year, though back in-line with the upwards trend from 2014 through 2020.
While the absolute number of rigs is lower, and often cited by those claiming that US production is going to decline precipitously, the new production per well drilled by each rig is much higher than 2018, and, the number of wells drilled per month, per rig is higher — all this amounts to a much higher amount of new production added annually, even with a lower rig count. Yes — this production works to offset declines, but with a US onshore decline rate of 31% (total, not for just new production), at 12mb/d, you need to replace around 3.75mb/d — so there is still slack in both the rig, and frac spread count. Working backwards, the rig count could fall another 10% (to 516), new production per oil rig (in the four major oil basins) fall to 1,050b/d, and wells drilled per month fall to 1.75 — and you would be adding around 3.7mb/d of production — so a slight decline, if things really start to slip — but oil rigs could fall to 450 if initial production improves 3%. So, continued scattered degradation in shale productivity can absolutely persist while US production is not under imminent threat of collapsing.
(Note, to calculate the total calendar weighted production per rig, you have to decline out each well drilled in a 12 month rolling period, then average the cumulative production over the period to get calendar weighted contribution. Individual basin decline is calculated as all wells on production before 2019, and the change in production from those wells through the end of 2019).
Not to forget, the Permian, and the US, has completely dwarfed global production growth relative to 2003, and especially so since 2013. It’s absolutely in the best interest of OPEC to regain control of the swing production. OPEC control will be a reoccurring theme this decade.
Recall these maps (from my April piece on the OPEC cuts), where the US has been the growth driver of the past two decades — such will no longer be the case going forward, and thus the US rig count becomes almost meaningless.
To put things in perspective, the size of the potential lost production, from the decline in US rig count, compared to global crude production — well, it’s minuscule. For one, NOCs, and super majors control ~60% of global production, compared to the US, which, including the Gulf of Mexico is ~15% (not a small number) — but when you consider that new well production from a single rig is ~1,000b/d (average, weighted across all basins), losing 40 rigs, is like losing a calendar weighted 150kb/d or — in reality, the rigs being dropped are marginal rigs, and it’s likely much less — probably closer to 130kb/d.
To put that in double perspective — Suncor’s corporate guidance range is 740kb/d to 770kb/d for 2023, CNRL has a 32kb/d guidance range, and Imperial Oil has a 20kb/d guidance range, and Cenovus has a guidance range of 40kb/d — together, the Canadian oil sands operators have a guidance range of 135kb/d — so a strong showing from the Ft. Mac boys could offset the loss from US rig drops YTD (and perhaps, a strong showing from the heavy boys is likely given the new TMX expansion).
At 1.5 wells per month, an average IP30 of 1,000b/d, and a decline rate of 10% per month (~60% first year decline), the 50 rigs we’ve lost from the plays they have dropped from is equal to a calendar average production per day of, in my estimation, ~122,500b/d for the year. Definitely not nothing, but it’s not changing the thesis significantly — the whole catalyst going forward for the oil trade, mostly hinges on a strong showing from China — if we’re being honest, any tightness created by US production declining, can be easily offset by the reversal of OPEC cuts should they feel like it.
There’s an interesting regression we can run, to put a number on the impact of US rigs dropping. See, in a bull market regime (you can decide if we are in one), the price of oil is highly sensitive to inventory draws — as plotted below, the correlation between “premium to implied inflation adjusted price” vs. the average rolling yearly inventory draw rate, per day — during the 2005 to 2007, and 2009 to 2014 bull market. The way I went about this, was, starting in 1993, when crude was $15, inflating the price of oil at 0.3% a month, then taking the premium at which spot trades to the “adjusted” price (see the top right chart). So, when we are in a bull market, and live pricing is sensitive to inventory movement, the value of a 100kb/d change in draw rate is around $2.00/bbl, so for a $20/bbl movement you need a 1mb/d movement in draw rate. So, yes, the change in US production isn’t nothing, but it might be worth only a few bucks on the oil price.
The remainder of this post is for paid subscribers, should you wish to subscribe, please follow the link below. Note — on June 1st, the price for new subscribers will increase to $55/month, or $500/year.
I am pleased to have added enterprise access to OilChem high frequency granular Chinese demand, refinery, and supply data which will significantly upgrade China crude and product coverage, the most important part of the crude story.
I have also added additional S&P Platts access, frac spread intelligence, live vessel tracking data — to improve the quality and breadth of my publishings, something I continually aim to do.
Existing subscribers will continue pay their current subscription price.