By Dennis Coyne
The labels in the chart above are for the "corrected (3 month)" estimate. This estimate (now preferred by Dean Fantazzini) is 35 kb/d lower than the EIA estimate for Sept 2016 (which is the most recent EIA estimate). Texas C+C output has fallen by about 500 kb/d from the April 2015 peak.
Dean Fantazzini said the following in an email:
I am still using a 3-month window because, after the structural break in February 2016, the correcting factors are still decreasing (even though at a lower pace). As discussed earlier, there are 2 reasons for this phenomenon:
1) digitalization at Texas RRC, which should speed up the processing of data filings,
2) lower production, which clearly results in a quicker data processing.
At the moment, unfortunately, it is not clear which factor is more important: in the medium/long term, the digitalization process at Texas RRC should make Texas data reporting similar to North Dakota and Pennsylvania where production data are almost final and only the last published month is usually subject to smaller revisions. However, it is not clear how long this transition will take. This is why I keep on using a 3-months average of the Texas vintage data.
For the past three months, the correction factors for C+C have been relatively stable, though potentially they may continue to fall so that RRC data will be relatively accurate after one or two months. For now, the Texas data needs to be "corrected" by 434 kb/d for the most recent month, with decreasing correction factors for earlier months (104 kb/d for 9 months earlier, Feb. 2016). A Chart with correction Factors over time below.
The chart above shows how the 3-month correction factors have changed over the past three months, a decrease for the most recent months (month 1 and 2) of about 150 kb/d over those three months and about 50 kb/d for month 6. The correction factors converge after month 12.
If the decrease in the correction factors continues linearly (which seems optimistic), they would fall to zero within about 18 months. We will continue to watch closely to see how quickly the Texas data improves.
The chart below shows how the "corrected" estimate using all vintage data has changed from July 2015 to Oct. 2016. It is fairly clear that the estimates were too high from about Jan. 2016 to Oct. 2016. The recent 3-month corrected estimates may be better, we will have to wait for more data in the future to know for sure.
Chart below has Texas C+C data from March 2014 to October 2016.
Natural gas data is below. A new 3-month vintage corrected estimate is shown, 22,316 BCF/d in Oct. 2016, a decrease of 398 BCF/d. The EIA reports 22,063 BCF/d in Oct. 2016, an increase of 258 BCF/d.