Musing: Marcellus Impact Study Rests on Some Shaky Assumptions

The latest study of the impact of the Marcellus shale formation on the economy of Pennsylvania continues the theme of prior reports by concluding that "…the development of the Pennsylvania Marcellus increased domestic energy production, creates jobs, and reduces government deficits." Earlier in its report, titled "The Pennsylvania Marcellus Natural Gas Industry: Status, Economic Impacts and Future Potential," the authors made the determination that if natural gas prices do not fall significantly in the future, "Marcellus economic activity could support over 250,000 jobs and generate $2 billion in annual state and local tax revenues." In a state beset by financial difficulties from a weak economy and lucrative state and local worker pension benefits, the prospect of a pot of $2 billion in new revenues has to be viewed positively.

This report, commissioned by the Marcellus Shale Coalition, is the third annual report a team of three energy-focused professors currently or formerly associated with Pennsylvania State University (PSU), have issued on the impact of the Marcellus shale on the state's economy. The Marcellus is considered one of the two hottest gas shale plays in the U.S., with the other the Eagle Ford shale in south Texas. In the PSU team's first assessment in 2009 when the Marcellus was early in its development, the professors compared the potential for the Pennsylvania resource to the largest gas shale basin, the Ft. Worth Basin's Barnett shale in north central Texas.  While the Barnett shale is still arguably the largest gas field in the United States, it will soon, if not already, relinquish the title to the Haynesville shale in Arkansas and Louisiana. 

The 2009 report contained a graphic showing how small the Barnett shale is compared to the Marcellus. The geographic spread of the Marcellus combined with its prolific wells led the PSU team to project it becoming the largest gas field in the country. As current production in Pennsylvania now exceeds local gas consumption, Pennsylvania has now become a gas-exporting state. And based on current drilling and completion activity, it would appear that this shift will not be reversed at any time in the foreseeable future, baring a total collapse of drilling.

Exhibit 1.  Barnett Shale Compared To Marcellus

Source:  Marcellus Shale Coalition

According to a 2009 XTO Energy investor presentation contrasting the similarities and differences between the Barnett and Marcellus shale formations there are 34 million acres in the Barnett with just two million of them in the core area of the basin. This contrasts with somewhere between 16 million and 32 million core acres in the Marcellus formation. The issue is whether all these millions of core acres in the Marcellus are truly core. 

Exhibit 2.  Successful Barnett Is Small Area

Source:  XTO Investor Presentation

Exhibit 3.  Core Marcellus Area Larger Than Barnett

Source:  XTO Investor Presentation

It is within the forecasting of drilling activity and well performance, both of which are responsive to estimates of future natural gas prices, where we begin questioning the gas production forecast in this latest report. In the 2009 report, the professors used natural gas price forecasts of $5.40 per thousand cubic feet (Mcf) for 2009 and then rising to $6.70 per Mcf in 2010 and thereafter gradually increasing until they reach $7.50 per Mcf by 2020. Under that scenario and using a typical production decline curve from the Barnett shale, the study projected gas production rising from 170 million cubic feet per day (MMcf/d) to 550 MMcf/d in 2010. They further saw the industry making Pennsylvania self-sufficient in gas supply by 2012 with production of 1,800 MMcf/d. By 2015, production was projected to reach 2,900 MMcf/d and almost 4,000 MMcf/d in 2020. 

Exhibit 4.  2009 Marcellus Shale Production Forecast

Source:  Marcellus Shale Coalition

In the 2010 report, the gas price assumptions changed to $5.41 per Mcf for 2010 and then increasing to over $6.00 per Mcf by 2020, nearly $1.50 below the previous year's target. The projected 2010 price assumptions called for an average annual real rate of increase of two percent. This projected price scenario would support an increase in drilling from an estimated 2,500 wells in 2012 to 3,500 per year by 2020. As a result of this activity outlook and the assumption of an economical ultimate recovery (EUR) per well of 2.8 billion cubic feet (Bcf), Marcellus gas production was projected to potentially exceed 2.5 Bcf per day in 2011, going over 7.0 Bcf per day in 2015 and reaching 13.0 Bcf per day by 2020.

Exhibit 5.  2010 Marcellus Shale Production Forecast

Source:  Marcellus Shale Coalition

In examining the assumptions underlying the 2011 forecast, we find that the professors used inflation-adjusted prices in 2010 dollars that yield a price of $4.58 per Mcf in 2012 and gradually rising to $5.30 per Mcf in 2020, more than $2.00 per Mcf below the value forecast in 2009. As a result of the lower price forecast, drilling activity was lowered to an estimated 2,300 wells to be drilled in 2011 increasing slowly to just below 2,500 wells in 2020. While the new well forecast has moderated, the professors raised their EUR to 3.6 Bcf and then increasing it steadily over the forecast period to 4.6 Bcf by 2020 to reflect industry advances in recovery technology. The net result of these revised forecast assumptions is that gas production is estimated to grow more rapidly – from 3.4 Bcf per day in 2011 to 12.0 Bcf per day in 2015 and further to 17.5 Bcf per day in 2020. 

When we consider all three forecasts we note that 2020's targeted production increased from 3.5 Bcf per day in the 2009 study to 4.0 Bcf per day in the 2010 one and now to 17.5 Bcf per day. That is essentially a five-fold increase in 2020 production, which is largely driven by assumed technological advances in completing wells that boosts the EUR. The fact that the projected production did not increase significantly between the 2009 and 2010 studies, even though the latter study used an EUR of 2.8 Bcf per day, we have to assume that the professors were using an EUR of closer to 2.5 in the 2009 study. For the 2011 study, the professors escalate the EUR from 3.6 Bcf per day to 4.6 Bcf per day over the next decade. But they start with a EUR estimate that is nearly 30% greater than the EUR employed in their 2010 forecast. Does the data support that change? It is interesting that consulting firm, Rystad Energy, forecasts Marcellus production in 2020 of only 7.9 Bcf per day or 45% of what the professors' project. The correctness of each forecast will have a profound effect on natural gas markets and prices, producer profits and the economy of Pennsylvania.

Exhibit 6.  2011 Marcellus Production Forecast

Source:  Marcellus Shale Coalition

If we look at estimates of the EURs suggested by producers active in the Barnett shale versus the cumulative production data from wells, there is no close correlation. (Exhibit 7 below.) This comparative data has been collected by Art Berman, a critic of the economic analysis underlying gas shale development, but the data doesn't seem to support the claims of gas shale producers. 

To understand the significance of this data, one should examine production from the Barnett based on historic wells and not including newly drilled wells. As the chart in Exhibit 8 shows, if one excludes Barnett wells drilled in the past 12 months, total gas production

Exhibit 7.  Well Production Not Supporting EUR Claims

Source:  Art Berman

declines at a 44% annual rate. That decline rate is consistent with gas shale well production profiles, but the rate of decline highlights the problem gas shale producers will face when and if they slow down drilling new wells in the basin. Without significant new well drilling, gas production is at risk, but the flip side of that risk is a higher natural gas price.

Exhibit 8.  High Drilling Needed To Sustain Production

Source:  Art Berman; Robert Gray

Coupled with the increasing EUR and a slightly increasing new well count in the professors' latest forecast, is the belief that gas shale wells can be drilled throughout the basin. That assumption ignores the growing reality that the Marcellus formation, despite being extensive, is not equally productive throughout the region. The map of producing wells in Pennsylvania demonstrates that there are essentially two areas of the state with above-average well production – the first area is the three northeast counties of Tioga, Bradford and Susquehanna, and then there are the two southwest counties of Washington and Greene. But it is also interesting to note the large number and the concentration of non-Marcellus wells in the western region of the state.

Exhibit 9.  Marcellus Wells Centered In Two Areas

Source:  PA Dept. of Environmental Protection

This sweet spot nature of the Marcellus formation is borne out in the following two charts. One shows the cumulative production of Marcellus wells, which substantiates the prominence for the state's gas shale industry of the five counties of Pennsylvania mentioned above. (Preparing these charts is a challenge because the data is not reported by Pennsylvania regulators on a monthly basis.)

Exhibit 10.  Cumulative Production Shows Sweet Spots

Source:  Art Berman; Lynn Pittinger

The shape of the chart in Exhibit 10 is further supported by data used to construct the chart in Exhibit 11. In this latter chart, the production history of Marcellus wells taken from the Pennsylvania Department of Environmental Protection is compared against the estimated monthly production that fits a decline curve designed to yield a EUR of 4.2 Bcf. What it shows is that wells this prolific are not only concentrated in those five counties, but actually are more concentrated in two of the three northeastern and one of the two southwestern counties.

Exhibit 11.  Well Performance Supports Sweet Spots

Source:  Art Berman; Lynn Pittinger

The next three charts show the wells that have been drilled in Pennsylvania in 2010 and in the first half of 2011 differentiated by Marcellus and non-Marcellus classification. In 2010, the well split was essentially 50/50, but for the first half of 2011, the ratio was 70/30 in favor of Marcellus wells.

Exhibit 12.  2010 PA Wells Drilled By Well Type

Source:  PA Dept. of Environmental Protection

Exhibit 13.  2011 PA Wells Drilled By Well Type

Source:  PA Dept. of Environmental Protection

When we look at the well permitting activity in the first half of 2011, Marcellus wells exceed non-Marcellus wells by a ratio of 55/45. The permitting is also focused in the core areas where Marcellus and non-Marcellus production has been established. 

Exhibit 14.  Pennsylvania Well Permitting By Well Type