Topic: Energy — Oil
Jul. 7 2015, 1:39 PM ET
- by VF member bz1516 (2475 )
Understanding EIA data and how to use it now --
All the discussion and comments weíve seen recently about EIA data and its reliability finally got me to take a deep dive into it. This is an important issue because US crude supply trends will determine the investability of oil producers and oilfield service including frac sand companies in coming months. My mission focused primarily on understanding the methodology of EIA reporting particularly supply data, and once that was accomplished drawing inferences about the direction of supply and pricing in the US.
The first thing I learned, in terms of making forecasts from EIA data, is the primary documents should be the monthly supply documents. These run two months behind the latest weekly documents. The latest monthly documents were just released this week. The latest monthly document is April.
Each month there is an adjustment made to supply numbers. These adjustments are a plugged number. But not to worry. This is a sound procedure. The disposition of the crude supply is considered more reliable. So the total supply number is made to equal the more accurate disposition number. Disposition is the way crude supply is used each month with the major use being refinery input.
Weekly data is model driven. The weekly data must be considered as subject to meaningful revision. Its weekly change is probably right in order of magnitude, but nobody knows by how much.
There are other levels of adjustment as well. Monthly data is reviewed annually and adjustments are made to prior monthly numbers going back up to two years. Occasionally as we saw in May there are adjustments made even more rarely than annually.
Weekly data are never adjusted after data is released. That is one reason why it is best to use monthly data. It is more consistent. The more important reason is that monthly supply data is actual data based on state by state reporting.
I personally donít believe it is worth spending much time analyzing the weekly data. I am using the weekly supply data, but just taking the data at face value as it comes off the report. I use it to update the trend since the last monthly report two months earlier.
Analyzing weekly adjustments would be subject to more error than the model driven production number taken at face value. It may be useful for purveyors of analysis to use weekly adjustments to drive home the point that weekly data should not be trusted, but the value in forecasting is very limited. Itís a better trade off to use the accuracy of monthly data rather than the freshness of weekly data.
Many people who read the EIA reports may not realize that the weekly or monthly adjustments donít make up any part of the published production numbers. They are in addition to the field production and import numbers. In fact the adjustments number could represent error in either field production, imports, exports, stocks change or refinery inputs. Itís just more likely the error is in either imports or field production. However the EIA leaves it to others to draw conclusions about where they originated. By making some realistic assumptions about the adjustment term it is possible to adjust field production to get a better picture of whatís really happening on a monthly basis.
Month after month, adjustments usually indicate some increase in field production. Sometimes minor, sometimes they are of larger size. Occasionally about once a year they can point to decreases in production, usually very modestly. Once in a great while as occurred in March they can have a more significant negative impact. This may happen about once every five years. Anyone saying adjustments can reduce production, except in very rare instances doesnít understand the methodology of how these EIA numbers are prepared.
Letís look at some numbers. In the last three months available, Feb Mar and Apr, field production was 9.432mm bbls/d, 9.692mm bbls/d and 9.701mm bbls/d. These numbers can be further refined by eliminating GOM and Alaska numbers, which contain no shale. After all what is important to most investors is forecasting what will happen with shale. The numbers net of the GOM and Alaska, those areas having no shale, are now 7.505mm bbls/d, 7.774mm bbls/d and 7.680mm bbls/d. These numbers reflect production in areas of the US that are primarily shale. There is what looks like a moderate size peak in March.
However we still have not dealt with the monthly adjustments. After applying my own allocation formula to the EIA monthly adjustments for the same three months I now show production of 7.612mm bbls/d, 7.636mm bbls/d and 7.793mm bbls/d mm for the three months ending in April. This shows a fairly steady increase through April.
Since the April report, May and June weekly production numbers indicate a drop from April to June of ~100,000 bbls/d. If these weekly numbers hold in the monthlies we can say a very moderate decline started in May. This outlook is likely more bullish than the market currently perceives. However based on the monthly numbers alone itís certainly possible we can see a leveling or even a slight increase. So overall its probably bullish.