Pennsylvania's Independent Fiscal Office found the state had the lowest tax burden on natural gas production out of 11 states studied, with Ohio just slightly above the Keystone state.
Headlines across Pennsylvania and in industry publications focused consistently on that particular conclusion.
Yet the comparison is not quite that simple.
The methodology for the 72-page report released late last week lays out how the IFO was able to compare Pennsylvania to states with quite different tax rules -- and why the office found it necessary to create its own metric, which the IFO called the "effective tax rate."
"We would encourage people to read the methodology carefully," the IFO's director, Matthew Knittel, said in an interview Friday.
Mark Ryan, the deputy director of the office, said that methodology is essentially a plea to consider the report only as it was conceived.
"The caveats are important to address what the analysis is intended to do and not intended to do," he said.
The report is intended only as an estimate of the future tax considerations relating to one Pennsylvania well.
"It is not designed as an annual effective tax measure," he said.
A letter attached to the report is addressed to state Sen. David Argall, the Schuylkill County Republican who requested it.
The report begins with a set of assumptions for determining what the taxes apply to: one hypothetical unconventional Pennsylvania gas well beginning production in a tight gas or other impermeable formation as of Jan. 1, 2014. The well is horizontally drilled and hydraulically fractured. The well produces only dry gas and only from a shale formation.
The 10 other states considered in the report are Arkansas, Colorado, Louisiana, Michigan, North Dakota, Ohio, Oklahoma, Texas, Virginia and West Virginia. Mr. Knittel and Mr. Ryan said they contacted each state to make sure their understanding of its tax model was accurate.
The report notes that Pennsylvania had the highest proportion of gas produced from shale of any state in 2012: 90.5 percent. Preliminary 2013 production data reported to the Pennsylvania Department of Environmental Protection suggest the state is second only to Texas in total withdrawals from shale by volume, according to the IFO report.
But in Texas, for example, only 45 percent of its extracted natural gas comes from shale formations, according to the U.S. Energy Information Administration. In addition, much of that is wet gas, containing other marketable compounds that increase the value of what's withdrawn.
Mr. Knittel said what was important for comparison's sake wasn't that the wells were identical, just that the report's model accounted for that particular state's tax rules.
In determining the amount of gas that the Pennsylvania well produces across its life, the report makes several assumptions.
First is that the IFO can determine what the producing life of the well actually is. Mr. Knittel said he assumed a 30-year productive life for the well because that seemed to be the industry standard. But that assumption is just that -- and thus imperfect.
"It is unclear what the exact productive life of the wells will be," he said.
No Marcellus well, for example, is anywhere near 30 years old, and new technology and techniques mean producers extract more and more gas at the beginning of a well's life. How that impacts a well's total production and its decline curve are not yet known.
The report's authors nonetheless believe their production estimate is close.
"Because such a large share of production occurs during the first 10 to 15 years of a well’s life, a shorter lifetime would not alter the general results of this analysis," according to the report.
A footnote addresses the report's estimation that the well would produce between 5 billion cubic feet and 10 Bcf in total, which the report's authors call "estimated ultimate recovery."
"Although the department publishes gas production data for January 2000 through December 2013, specific data related to natural gas extraction in the Marcellus shale are only available since July 2009," according to the footnote. "Hence, historical data that could be used to inform estimates of the EUR [estimated ultimate recovery] for a new well are limited."
The production estimate is higher than that from the federal Energy Information Administration because the Pennsylvania calculation takes into account recent drilling efficiencies, such as more fracking stages and longer laterals.
"This EUR range is generally consistent with very recent production data reported to the Pennsylvania DEP," according to the report.
"However, in order to project EURs, decline curve analysis requires the specification of three variables: initial production, initial decline rate and the long-term decline rate. Insufficient data exist to quantify the long-term decline rate (hyperbolic decline exponent), so that parameter is an estimate, and ultimate EURs in the Marcellus shale remain undetermined."
The report also addresses what its authors call "marketed production," which excludes gas used in production, venting and flaring. But another footnote calls some of that data into question.
"If data are limited," according to the footnote, "gross withdrawals may equal marketed production."
Mr. Knittel said he realized there is unlikely to be a case where every cubic foot of gas withdrawn from a well is sent to market. But he couldn't reconcile EIA data with what the companies reported to the separate states, so he had to make a choice to allow his model to consider gas brought to market.
Operators in Pennsylvania and elsewhere do vent, flare and use gas in production, Mr. Knittel said. But in the report, Pennsylvania's marketed production -- 2,256,696 MMcf, according to the report -- equals its gross withdrawals.
The same thing appears in the report for states explicitly studied and others included for reference, including Louisiana, Oklahoma, Colorado, West Virginia, Utah, Alabama, Virginia, Michigan, Kentucky, and Ohio.
For Texas, Wyoming, New Mexico, Alaska, and North Dakota, gross withdrawals are higher than marketed production because Mr. Knittel said he had better data, meaning he knew how much gas was flared, vented or used in production and thus not sold.
In determining market prices, the report uses the Henry Hub spot price, the price point for the New York Mercantile Exchange measured at a terminal on the Gulf Coast. That price tracks closely with national prices over time, according to the Energy Information Administration, but the IFO report notes that it is not a perfect substitute. In order to make any comparisons, however, the report's authors needed a constant price point.
"Gas prices may vary by geographic region, but the analysis must apply the same high and low price scenarios across all states to facilitate a meaningful interstate comparison," according to the report.
In calculating the price, the report deducts some post-production costs, but not all. It subtracts gathering and transportation costs from the marketable price, but not processing.
"We didn't including the processing because we assumed it was all dry gas," Mr. Knittel said.
Considering whether that dry gas was compressed into liquid natural gas or some other form was outside the scope of the report, he said.
Having estimated production and price, the report moves on to taxes.
The IFO's metric, the effective tax rate, was applied to four scenarios: high and low prices for a high-producing well and high and low prices for a low-producing well.
In each case, Pennsylvania's came in at the lowest rate with Ohio close behind.
A low-producing Pennsylvania well would be taxed at an effective rate of 1.6 percent with low prices and 1.3 percent with high prices. The effective tax rate for a high-producing Pennsylvania well would be 0.8 percent with low prices and 0.6 percent with high prices.
But in creating a single tax rate, the IFO had to account for differences between the states' tax structures.
Pennsylvania is the only state of the 11 without a severance tax, using the impact fee created by Act 13 to collect from operators. And states with severance taxes calculate them differently and have varying provisions for collecting other taxes.
Gas reserves are exempt from real property taxes in some states, including Pennsylvania, but taxable in other states. In Arkansas, they're assessed at 20 percent of full market value. In Ohio, they're assessed as a percentage against the whole property assessment and can either be raised or lowered county by county. In Colorado, they're taxed at 87.5 percent of the gas sales price for the preceding year.
The report also cannot account for all of the taxes that may follow production in various states, such as income taxes.
"It is much more difficult to quantify any income taxes that would be levied on the profits from extraction," according to the report. "In order to quantify those amounts, the analysis must specify the profit margin on total sales. The analysis must then identify the type of entity to which the income ultimately accrues.
"If the entity is a corporation and subject to the corporate net income tax, the relevant apportionment formula and factor must be specified. Many large natural gas extractors are multistate corporations that have unique factors to apportion firm-wide income to the states in which they operate. Hence, even if total profits could be identified, it is unclear how much would be taxed at a particular state corporate income tax rate."
Mr. Knittel said explaining the limitations of the report was important so its findings wouldn't be misused. The IFO is barred from making policy recommendations, he said, and sought only to build a model to examine what it considered a necessary set of data.
"I can't think of a better way to do it, to make it comparable," he said.
Jacob Quinn Sanders: firstname.lastname@example.org