- The EIA drilling productivity report has been used to estimate decline rates in the Bakken, Eagle Ford and Permian light tight oil (LTO or shale) plays (Figure 1). The objective is to estimate how far production will fall in these plays in light of the sharp decline in US rig count. The rationale is to calculate the production level at which new oil production capacity added will cancel the declines and equilibrium is reached.
- The shale drilling industry is highly dynamic. While rig count has fallen by about 60% in the 3 plays since end 2014 this is in part offset by rig productivity that has more than doubled in recent years. The rig count decline has more recently stabilised. Improved productivity should help profitability, but this is cancelled by the collapse in oil price.
- The estimated recent annual decline rates are 47% for the Bakken, 55% for the Eagle Ford and 22% for the Permian. These declines are not constant. Since year one declines are often of the order 70% and the smaller number of wells now being drilled means that the number of fast declining year one wells in the production pool is falling. Play decline rates are therefore also falling with time.
- With these declines and current rig count and productivity levels, production in the Bakken will stabilise at around 870,000 bpd, down 330,000 bpd on current levels. Production in the Eagle Ford will stabilise at around 1,140,000 bpd, down 560,000 on current levels. The lower decline in the Permian means that production there will continue to rise. It is estimated that the net effect will be an LTO production decline of the order 830,000 bpd spread over several months.
- It is questionable whether a decline in LTO production on this scale will be sufficient to bolster the flagging oil price and may, for example, be offset by production gains in Iran and elsewhere. Since much of the global oil industry cannot survive at current price levels a second and more brutal round of cuts to OECD companies is to be expected. This may in part take the form of company insolvencies that are just getting underway in the US shale industry. Ultimately, a balance between global oil supply and demand must be restored and, without a production cut by OPEC, this must then occur within the non-OPEC companies and countries.
Figure 1 The main shale oil and gas plays of the USA 
The Energy Information Agency (EIA) of the USA publishes a monthly Drilling Productivity Report detailing the production and drilling statistics for the 7 shale oil and gas regions of the USA (Figure 1). This post presents an analysis of this data for the Bakken, Permian and Eagle Ford, that combined, account for 89% of US shale oil production. Much of this production does not actually come from shale but from “tight” formations that need to be fracked to allow the oil and gas to flow. Shale oil has therefore been re-christened as light tight oil (LTO). In this post I use both terms – they both mean the same thing.
At this point I need to add a disclaimer. There are reasons to doubt the reliability of all the data in the EIA drilling productivity report. For example some have questioned whether all the Permian wells counted are in fact LTO wells. Some may be in more conventional oil pools. The interpretations of the data presented here can only be as good as the input data upon which they are based.
For each region, for oil and gas, the EIA publishes 3 key statistics: 1) monthly production, 2) number of operational rigs 3) production added per operational rig. This data allows for an analysis of the decline rates to be made. Declines in shale are known to be extremely high, often 70% in the first year, and many estimates of play declines have already been made. Because of the collapse in US oil directed drilling, shale oil production was also supposed to collapse sending the oil price back up. Since this has NOT YET happened, I decided to take a look at the numbers myself.
If no new wells are drilled in a shale region with 50% decline rate then at the end of a year production will halve. Declines are not constant over time and tend to reduce in maturing wells. Hence production will fall by less than 50% in the second year.
In the past, new and “frenzied” drilling in the US shale patch has done three things: 1) it has compensated for declines, 2) it has added to production and 3) a certain number of wells have been drilled and not completed. Hence there is / was a stock of drilled and uncompleted wells and this makes the analysis presented here more complex and less certain. Generally there is a problem matching drilling one month directly to a production change that month. This has been partially overcome by smoothing the data.
The logic applied is that the difference between production added and production change each month equals the production decline. For both production change and production added I have used a 7 month centred moving average to calculate declines (see charts below). Information on drilled but not completed wells is not readily available and has been ignored. There is probably a better way of doing this but the results I get are similar to those produced by others. For example, Ron Paterson estimated 54% annual decline for Bakken and 62% for Eagle Ford in May of this year.
For each of the Bakken, Permian and Eagle Ford areas 5 standard charts have been produced. These charts are presented below with key observations and the discussion follows.
Figure 2 Raw data for the Bakken showing monthly production and rig count. Note how rig count peaked in June 2012 but that production has continued to rise steeply. The recent peak in September 2014 followed by a sharp decline in rig count has resulted in a production plateau. There are four reasons for this: 1) time lags between drilling, completion and first production, 2) a substantial improvement in rig productivity (Figure 3), 3) there are still 75 rigs drilling that helps compensate for declines and 4) old drilled but uncompleted wells are perhaps now coming on line (an accentuated version of reason “1”).
Figure 3 Drilling productivity as reported by the EIA based on their analysis of available data. Note how back in January 2011 a single rig added on average 200 bpd new production capacity. By mid-2015 that number has risen to over 600 bpd, a three fold improvement in efficiency in less than 5 years. Note there are 3 variables that can combine to improve efficiency: 1) drilling faster (more wells / rig), 2) drilling and fracking better – improved completion technology and 3) targeting more productive zones.
Figure 4 By combining the rig and productivity data shown in Figure 4 we can produce this chart that shows production capacity added each month by multiplying number of rigs by the production added per rig. Not surprisingly, production added has collapsed along with the rig count. But almost 50,000 bpd per month is still being added in the Bakken. Is that enough to cancel declines?
Figure 5 The monthly production increase comes from Figure 2 and the monthly production added from Figure 4. Note that both curves have a 7 month centred moving average applied. The difference between production added and production increase is accounted for by decline (Figure 6).
Figure 6 The difference between production added and production increase is accounted for by decline (Figure 5) converted to a monthly % of the production in the prior month. The 7 month smoothed data (dark blue line) conveys decline rates that vary from 3 to 8% and a tendency for annual cycles. This variability in calculated declines almost certainly reflects in part operational dynamics. The 13 month smoothed line provides a better picture of the underlying trend. Decline rate increased when the 2009 drilling frenzy got under way – larger numbers of fast declining year one wells and has stabilised at around 5.2% per month. Note that annual decline is not 12*5.2% but is rather 47% (calculated from spreadsheet).
The Eagle Ford
Figure 7 The Eagle Ford rig count and production profiles are very similar to The Bakken (Figure 2). The Eagle Ford rig count peaked in May 2012 and has since more or less moved sideways until the late 2014 crash. Eagle Ford production began to decline in April 2015. Note that at 1.7 Mbpd, Eagle Ford production is substantially higher than Bakken.
Figure 8 The improvement in rig productivity for the Eagle Ford is even more impressive than for the Bakken rising from below 100 bpd / rig / month to over 700. Productivity continues to rise post-crash.
Figure 9 As with the Bakken, Eagle Ford ‘production added’ has plunged, following the rig count down. But 80,000 new barrels per day per month are still being added to capacity.
Figure 10 The monthly production increase comes from Figure 7 and the monthly ‘production added’ from Figure 9. Note that both curves have a 7 month centred moving average applied. The difference between production added and production change is accounted for by decline (Figure 11). Note how production change in the Eagle Ford has now turned negative and production has begun to decline (Figure 7).
Figure 11 The Eagle Ford decline rate rose from 4 to about 8% per month but is now clearly falling as the number of high decline year 1 wells is reduced. Using the more recent number of 6.5% per month yields an annual decline of 55%. Using the prior 8% per month figure yields an annual decline of 63%. The new well productivity of the Eagle Ford is the highest of the three regions but so are the declines.
Figure 12 The Permian is the USA’s largest LTO producer with over 2 Mbpd production. And while the Eagle Ford had a standing start in 2009, the Permian already had 850,000 bpd in 2007. The 2008 crash somewhat surprisingly does not show up significantly in the production data. One possible explanation is that pre-2009 production was dominated by conventional wells that would not be so prone to rapid decline. The rig count pattern is similar to the other areas, but notably rose to a new high in 2014. The Permian is clearly preferred by operators to the Bakken and Eagle Ford plays.
Figure 13 The rig productivity in the Permian has also risen significantly. But the rig productivity in the Permian (220 bbls per rig per month) is markedly lower than in the Bakken (over 600) and the Eagle Ford (over 700). The Permian also has a much higher rig count. It also shows a sharp and strange uplift in very recent rig productivity. At this point I can only speculate that production has received a boost from some “old” drilled wells being brought on line. So, if the Permian has the lowest rig productivity, why is it the most popular and largest play? The answer to that lies in the decline rate.
Figure 14 Production added does in fact follow rig count quite closely but has turned up in recent months due to the “surge in rig productivity”. Production has begun to rise even with the rig count still trending down.
Figure 15 The production dynamics in the Permian are much more favourable than in the Bakken and Eagle Ford. While production change in the Bakken is now zero, has turned negative in Eagle Ford, it is still rising in the Permian, the largest play. This will have profound impact on the aggregate behaviour of the US shale oil patch.
Figure 16 Decline rate in the Permian appears to be continuously variable and subsequently hard to estimate. It peaked at over 5% but has since fallen to about 2.5% per month. The slowdown in the decline rate will again represent the smaller number of fast declining year one wells in the production pool. 2.5% monthly decline translates to an annual decline of 22%, well below the Eagle Ford and Bakken. The 230 rigs currently drilling are able to more than compensate for the declines, hence Permian production is still rising.
With the Bakken, Eagle Ford and Permian we are dealing with about 89% of US LTO production. The production dynamic differs substantially between the three plays as summarised in Figure 17. Hence it is difficult to make general statements about US shale oil production. Observations made today will be different in 6 months time. In particular, decline rates vary a lot with time making prediction difficult.
Figure 17 Summary production and drilling data for the main US shale oil plays.
In the Eagle Ford and Permian, decline rates are tending to fall as the number of fast declining year one wells reduces in the production pool. Hence what I have to say now based on most recent declines may prove to over-estimate the actual decline in US LTO production. The annualised most recent declines are as follows:
- Bakken: 47%
- Eagle Ford: 55%
- Permian: 22%
The other key statistic is that the oil directed rig count in the USA has stabilised, currently 31 oil directed rigs higher than the low point reached on 26th June. IF we continue with the current rig count, rig productivity and declines then Bakken production will stabilise at about 870,000 bpd, down 330,000 bpd on current (in 6 months time).
This is calculated by taking recent production of 1,200,000 bpd and applying a 5.2% decline per month and observing that this will take production down to 871,000 bpd after month 6. At that production level the Bakken lost 48,000 bpd between months 5 and 6 equal to the amount of production added in June 2015. That is the equilibrium point where applying current declines and drilling dynamics results in production added cancelling declines. The calculation is subject to many assumptions and uncertainties.
Eagle Ford production will stabilise at around 1,140,000 bpd, down about 560,000 bpd on current (in 6 months time). Permian production will continue to rise by perhaps 10,000 bpd per month (60,000 bpd in 6 months time). Then, a new equilibrium will be reached in the shale patch where drilling compensates for declines at a production level very roughly about 830,000 bpd below current levels. If declines continue to fall, productivity continues to rise and rig count starts to rise then the production fall will be less (whilst noting that more new wells drilled will accelerate the aggregate decline rate).
So is this good news? With US production riding high at 12,910,000 bpd I suggest that the rout in the shale patch will merely take this back towards 12 Mbpd. And then we need to consider what is happening in the conventional US production portfolio. Current 12.9 Mbpd production is split roughly 5.4 Mbpd in shale and 7.5 Mbpd conventional. In the North Sea the record high price of the last 5 years is feeding through to future production increases. The same may happen in the USA.
Hence, the main message from this post is that a precipitous fall in US production in the months ahead, upon which most analysts are depending upon to send the oil price higher, may not materialise YET. This is simply the end of round one of the current oil price crisis and the standoff between US shale and OPEC.
Is it good news or bad news that US oil production may not collapse (yet) under the weight of low oil price? It’s certainly good news for US energy security. And if US production does not collapse, it will in the short to medium term likely lead to a further decline in the oil price, great news for consumers and the economy as a whole.
The picture from the producer side of the fence is rather different. Sub-$60 will mean that many conventional oil producers are in great difficulty. Many shale producers are arguably already insolvent. High debt levels are secured against assets, i.e. oil and gas reserves. Reserves in turn vary with oil price and as the price goes down so arguably does the volume and value of those reserves. It is widely believed that for the oil price to recover, OECD companies must pump less oil. Pumping less oil for lower price will be route 1 to bankruptcy for many highly leveraged companies producing high cost oil. I believe the industry may face a crisis unparalleled in recent decades.
The future is just about to begin.
 EIA: Drilling Productivity Report