Crude oil and natural gas production growth stalled in 2015 and has declined this year in some of the big shale basins. But we may be seeing a turnaround. The latest EIA Drilling Productivity Report, released on December 12, 2016, included upward revisions to its recent shale production estimates and also projects an increase in its one-month outlook for the first time in 21 months (since its March 2015 report). Today we break down the latest DPR data.
Since its inception in late 2013 (we first introduced it to the RBN blogosphere in Higher and Higher), the DPR has become an industry bellwether for production trends. The report provides a leading indication of expected crude and natural gas production from seven leading shale basins across the U.S. To do that, EIA takes its historical monthly production estimates (which are lagged by two months), well production data from individual states as well as rig counts and other data sources and estimates historical and recent rig productivity. It then comes up with implied oil and gas production volumes for each month going forward, including the two previous months where EIA data is not yet available, as well as the current month and one forward month. It does this for the seven major U.S. shale plays: Bakken, Eagle Ford, Haynesville, Marcellus, Niobrara, Permian and Utica. We provided a detailed explanation of the DPR’s model inputs, methodology, assumptions and risks in our blog series “Every Rig You Take.” But in short, the aim of the report is to estimate the net impact of natural declines from existing wells (older than 30 days), productivity changes and new drilling on production volumes on a monthly basis. The bottom line is if improving rig productivity is more than offsetting natural declines or slower drilling activity, then production is expected to increase. If the decline in volume from existing wells is the larger of the two, then DPR shows a decline in production. As another month of actual production volumes and other inputs become available, the DPR revises its historical and forward estimates in the next release.
The latest DPR not only shows upward revisions in recent production estimates (for both oil and gas) but also indicates a potential turning point in production trends––for the first time since the March 2015 report, the DPR is projecting a net increase in production in January. That is not to say we have not seen actual month-on-month increases in the last 21 months—we have. In many months, as more data became available and is applied retroactively, the DPR’s projected declines have turned into actual increases. However, the March 2015 report was the last time the DPR’s analysis of drilling activity and rig productivity indicated that production would increase. So let’s take a closer look at the data, starting with what’s changed in DPR’s latest oil and gas estimates. Figure 1 compares the actual and implied estimates for total shale oil (left graph) and shale gas production (right graph) from the latest (December 2016) report and the previous (November 2016) report. The solid lines in the graphs are considered more or less actuals (likely to see lesser revisions), while the dashed continuation lines are implied estimates based on the model inputs described above.
Figure 1; Source: EIA
The December DPR data revised its estimates for shale oil upward an average of about 45 thousand barrels per day (Mb/d) for May 2016 to December 2016 estimates, as seen by the divergence between the blue (December 2016 data) and gray (November 2016 data) lines in the left graph (purple circle). Both month’s datasets show a steady decline in shale oil production starting in early 2015, with few exceptions, and continuing through December 2016. Similarly on the gas side, the December DPR revised its estimates for shale gas up an average of nearly 350 million cubic feet per day (MMcf/d) for August 2016 forward (red circle), with smaller upward revisions prior to August. Overall, the December data continues to show shale gas production has declined, by nearly 0.5 Bcf/d between August and December (2016).
However, in the case of both commodities, with the addition of the January 2017 outlook in the December report, the DPR is showing the first net increase in its month-ahead projection since its March 2015 report. For instance, Figure 2 shows the projected month-ahead increase or decrease in shale gas production from each month’s DPR report since the report began publishing in November 2013. Until that March 2015 report, the DPR model inputs had been consistently predicting that gas production in the next month will increase from the current month. In the February 2015 report, the model showed the smallest projected increase since the report began publication. And in the March 2015 report, the model projected the first month-ahead production decline since late 2013. Each monthly update since then has predicted a month-ahead decline—until now. As of the December 2016 data, the model implies that shale oil production will edge up about 1.4 Mb/d in January 2017, while on the gas side, shale production is expected to tick up a net 88 MMcf/d. That is to say that, after steady declines in recent months, current rig counts and rig productivity indicate production may regain some footing in the new year and as early as January 2017, if DPR’s estimates materialize EIA expects.
Figure 2; Source: EIA, RBN
Granted, the projected increases in total production aren’t substantial, especially compared to the monthly growth that used to occur prior to 2015. And the revised December data shows that larger month-on-month increases than that have occurred in recent history. However, it may be an early signal of the timing of a potential (albeit modest) turning point in production trends.
As per the DPR methodology, there two main factors underlying this shift: rig counts and rig productivity. The DPR shows a total 139 rigs were added to the active list between June 2016 and November 2016, over 60% of those (85 rigs) in the Permian, a combined 13 in the Marcellus/Utica, 12 in Niobrara, 11 in the Bakken and seven each in Eagle Ford and Haynesville. In addition, the latest Baker Hughes weekly rig count data shows the total U.S. count is up another ~60 combined since the end of November. So rig counts are up, and so is rig productivity (new well production per rig), as shown in Figure 3, which compares the implied new-well production per rig expected in January 2017 by play (dark brown columns for oil and darker blue for gas) versus what it was a year earlier in January 2016 (light brown for oil and light blue for gas). With just one exception (new oil wells in Utica, which is mostly condensate anyway), production per rig is notably higher across the board. In other words, not only are rigs returning, but they are presumably returning at a higher efficiency than when they left.
Figure 3; Source: EIA
It’s worth noting here that historically productivity has a tendency to decline as rig counts increase. This happens because more rigs can mean more diversified rig placement in potentially less productive areas. But as the folks at EIA pointed out to us last week, that may not happen this time around, or at least not to the extent it has in the past. In particular, they referred us to the graph in Figure 4 below showing historical rig counts (black line using the right axis) and rig productivity (brown line using the left axis) for the oil-focused Bakken play.
Figure 4; Source: EIA
As the graph shows, through much of 2009 (red circle in Figure 4), the rig productivity increased as rig counts declined (and remaining rigs likely were focused on “sweet spots”). But then as rigs returned later in 2009 and into 2010, productivity declined. This year, the recent and projected productivity trend (brown line in gray oval on the right) does not reflect that same response in productivity to rising rig counts—at least not yet. In speaking with the folks at EIA about this, they acknowledged that it’s too early to tell for sure what if any impact rig counts will have on rig productivity in late 2016 and in 2017 (we’ll be watching the monthly DPR updates for that). But they noted that there aren’t any signs of declining productivity just yet, and moreover, that the latest state-level data from North Dakota suggests that some of the experimentation with “high sand loading” (an enhanced well completion technique) may bring positive results and we may not see significant reduction in the productivity numbers. Other reasons that may stave off productivity declines: the availability of ample drilled-but-uncompleted wells (DUCs; see DUC, DUC, Produce) and, simply, better rigs than what were available 6-7 years ago. In the Bakken, for instance, EIA said it estimates there were fewer than 100 DUCs back in 2010. Now according to EIA, there are over 800 DUCs in the Bakken that can theoretically be completed and in effect skew productivity rates higher.
All that said, there are some other risk factors to keep in mind as well. On the upside, the DPR does not take into account pipeline expansions that may allow more production to occur, specifically gas out of the constrained Marcellus/Utica. On the downside, the DPR also doesn’t account for weather or maintenance-related impacts on production, such as freeze-offs or temporary capacity outages that may restrict production (see Cold As Ice). And, finally, it doesn’t predict the potential impacts on production of changing economics in a play. Any one of these factors could end up proving the DPR’s January 2017 estimate wrong.
It should be noted here too that while the DPR includes estimates for most of the major oil and gas shale basins—where the activity and growth is happening—it does not include non-shale areas, where more than half of all U.S. crude oil and about 40% of gas production still comes from. Most of that non-shale production is flat or declining. So an increase in shale production does not necessarily translate into higher overall U.S. production volumes. For that to happen, the growth from the shale areas must be enough to more than offset any declines from those other areas. The current rig additions and productivity improvements suggest there will be a better chance of that happening in 2017 than there was in 2016, particularly on the gas side. Thus, we’ll be watching for revisions and trends in the next DPR, to be released January 17, 2017. And of course, we closely track U.S. natural gas production on a daily basis in our NATGAS Billboard pdf report and data file based on daily pipeline flow data.
“High Hopes” was written by Jimmy Van Heusen with lyrics by Sammy Cahn and was recorded by Frank Sinatra in 1959 in a hit version from his 1961 album, All the Way. The song won an Oscar for Best Original Song at the 32nd Academy Awards in 1960 and reached #30 on the Billboard Hot 100. Completely different songs of the same name were sung by Pink Floyd, Bruce Springsteen and Sammy Hagar.