Edit 12/2/2016, 8:30 AM EST. A corrected chart for correction factors below and a new chart added.
The railroad commission of Texas released September output data a few weeks ago and Dean Fantazzini made an estimate of the “corrected” data for crude plus condensate (C+C) and natural gas. Last month he found there had been a structural change in the data in March and provided a new estimate using only the most recent 3 months of data, I chose to use the most recent 6 months of data instead.
This month I will present the old estimate (labelled “all”) using all vintage data (dating back to April 2014) as well as the 6 month and 3 month estimate, which use the most recent 6 months and 3 months of data respectively.
To clarify, the data reported in November, October, and September for the most recent 48 months in each case (144 data points in all) are used to determine the correction factors for the most recent 24 months for the “3 month” estimate. Likewise 6 sets of data are used for the “6 month” estimate, and 30 sets of data are used for the “all” estimate. In each case the correction factor is the average of the 3, 6, or 30 sets of correction factors for each of the most recent 24 months (Oct 2014 to Sept 2016).
Dr. Fantazzini prefers the 3 month estimate, so in the future I will use this rather than the 6 month estimate, last month the 3 and 6 month estimates were very similar and that remains the case this month.
In September 2016 Texas C+C output fell by 38 kb/d to 3236 kb/d, based on the 3 month estimate, the EIA estimates that output rose by 6 kb/d to 3163 kb/d.
The chart below looks at the natural log of TX C+C output to estimate the annual decline rate since the peak in March 2015,the average annual % decline rate is the slope of the trend line times 100.
For the “3 month” estimate the annual decline rate is 6.31%/year and for the EIA estimate it is 8.75%/year since March 2015.
The chart above shows how the correction factors have changed from April 2014 to Sept 2016, I only show the correction factors for the most recent 9 months, the correction factors get progressively smaller and less variable for earlier months (10 through 23). The vertical axis is in barrels per day, divide by 1000 for kb/d, so the highest correction factor in April 2015 is about 1000 kb/d for the most recent month (1 on the chart).
The correction factors for the most recent 5 months has seen a marked drop since Feb 2016, though a similar drop was seen in April and May of 2014, maybe the RRC data has improved, we will see in the future. The estimates for C+C output are very similar from 10 months and earlier (before Jan 2016) whether the “all”, “3 month”, or “6 month” estimate is used.
The chart above compares 4 different estimates from June 2015 to Sept 2016, all, 3 month, 6 month, and 12 month average correction factors where the most recent 3, 6 and 12 months of correction factors are averaged, and “all” uses 30 months of correction factors based on RRC data from April 2014 to Sept 2016 (1440 individual data points).
The corrected Texas Natural Gas estimate is shown in the chart above, based on Dr. Fantazzini’s estimate there was a 1.14 million CF per day drop in Texas Natural Gas output in Sept 2016 to 22.7 MMCF/d. The EIA estimate for Sept 2016 is 21.8 MMCF/d. Note that in the past there have been problems with the natural gas data from the RRC, so the big drop from August to September might be a data anomaly.
The chart that follows shows the natural log of Texas Natural Gas Output. The slope of the trend line times 100 is the annual decline rate in per cent. Since May 2015, the corrected estimate has an average annual decline rate of 4.1%/year and the EIA estimate declines at 8.9%/year.