Why EIA's Crude Production Stats Are Flawed

Summary

• Shale oil economics are different than conventional oil.
• EIA developed a model to try to estimate its production.
• But the industry is managing drilled wells strategically.
• That makes EIA’s model inherently flawed.
• EIA has been underestimating production.

The Energy Information Administration (EIA) does not collect current survey information in estimating the weekly U.S. crude production it reports in its Weekly Petroleum Status Report (WSPR). Instead, it relies on modeling.

With the rapid development of shale oil reserves in the U.S., the EIA had to develop a new model to estimate output because shale oil economics and production are so different than conventional resource development.

Drilling Productivity Report (DPR) Model

The DPR model is a computer model that estimates recent and future and oil and gas production for seven specific regions (DPR Regions) in the U.S. The DPR model parameters are based on one single input: Baker Hughes drilling rigs counts. By regressing rig counts and monthly production, it estimates a relationship between the two.

DPR Regions. Source: Energy Information Administration.

The DPR model is re-calibrated each month to estimated how much oil and natural gas production in a play has increased as a result of the drilling and completion of new wells, and how much production from existing wells has declined as the resource in place for the existing wells is depleted (click here for detailed methodology). The DPR model also estimates the time delay between the start of the drilling of a well and that well's initial production.

The DPR model was designed to provide estimates of recent historical oil and gas production and projections of future oil and gas production on a monthly (not weekly) basis. The model parameters are re-estimated each month to reflect the long-term technological trends regarding the configuration of the drilling rig fleet, the rate at which new wells are drilled and completed, and the design of new tight oil and shale gas wells. All three trends collectively change the rig efficiency and the well productivity.

The DPR model was designed to project short-term oil and natural gas production volumes for seven of the DPR regions as input to EIA's Short-Term Energy Outlook (STEO) projections for total domestic oil and gas production. But one major limitation is that rig counts are used from two months prior to the production month, and so EIA must forecast future rig counts to get a forecast more than two months out.

The DPR model is also used to develop estimates of current and recent historic oil production, as reported in EIA's Weekly Petroleum Status Report (WPSR). The EIA explains that the estimated change in production does not reflect external circumstances that can affect the actual rates, such as infrastructure constraints, bad weather, or shut-ins based on environmental or economic issues. The biggest obstacle with this approach is that oil producers have developed a large inventory of drilled but uncompleted wells (DUCs) that may have be affecting actual production or may affect production rates in the future.

However, according to Baker Hughes, "To be counted as active a rig must be on location and be drilling or 'turning to the right'. A rig is considered active from the moment the well is "spudded" until it reaches target depth or "TD". Rigs that are in transit from one location to another, rigging up or being used in non-drilling activities such as workovers, completions or production testing, are NOT counted as active." This means that the DPR model is not built to assess production from DUC wells that are being completed on a delayed basis.

As discussed in more detail in my previous article, the inventory of DUCs is estimated at around 4,000, which may be capable of bringing 500,000 barrels per day on-line quickly, should companies pull the trigger. If prices rise, shale producers will be incentivized to achieve a higher ROI. If prices remain depressed, operators may need to liquidate inventory to maintain cash flow. Regardless of prices, North Dakota has given operators another year to commence production from drilled wells, and the prospect of a total loss on DUCs is unacceptable if prices are high enough to justify completion costs.

As a result, the EIA cannot estimate the production from DUCs from regression analysis because the operators' behavior has not has much of a history to model. Furthermore, as noted above, any rigs that are involved in completions do not appear in the rig counts.

Comparing the weekly estimates with EIA's monthly survey numbers, shows that the EIA estimation procedure has mainly underestimated production. The biggest exception is June, when EIA realized it was understating production for February and March and upwardly revised its estimate in one fell swoop:

MBD	Monthly	Weekly	Diff
Jan-2015	9345	9180	-165
Feb-2015	9456	9278	-178
Mar-2015	9653	9398	-255
Apr-2015	9694	9381	-313
May-2015	9479	9431	-48
Jun-2015	9315	9599	284
Jul-2015	9433	9520	87
Aug-2015	9407	9324	-83
Sep-2015	9460	9121	-339
Oct-2015	9347	9126	-221

Click to enlarge

Note: Monthly data source is the EIA Petroleum Supply Monthly; weekly data source is the EIA Weekly Petroleum Status Report.

Conclusions

Shale oil production presented a new modeling challenge for the EIA. Although it developed a reasonable approach, the industry started to manage its DUCs as a strategic asset.

There is a large inventory of DUCs which can have a material effect on U.S. production. At this point, it is difficult to predict the timing of its impact, but there is reason to think it could hit 2016 numbers due to regulatory issues as well as survival issues. If prices were to advance much, it is a safe bet that a lot of production will materialize from DUCs.

Analysts should not quote weekly changes in EIA production statistics as fact. It is a guesstimate from EIA's monthly model.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.