Fracking's Secret Problem - Oil Wells Aren't Producing as Much as Forecast
Thousands of shale wells drilled in the last five years are pumping less oil and gas than their owners forecast to investors, raising questions about the strength and profitability of the fracking boom that turned the U.S. into an oil superpower.
The Wall Street Journal compared the well-productivity estimates that top shale-oil companies gave investors to projections from third parties about how much oil and gas the wells are now on track to pump over their lives, based on public data of how they have performed to date.
Two-thirds of projections made by the fracking companies between 2014 and 2017 in America’s four hottest drilling regions appear to have been overly optimistic, according to the analysis of some 16,000 wells operated by 29 of the biggest producers in oil basins in Texas and North Dakota.
Collectively, the companies that made projections are on track to pump nearly 10% less oil and gas than they forecast for those areas, according to the analysis of data from Rystad Energy AS, an energy consulting firm. That is the equivalent of almost one billion barrels of oil and gas over 30 years, worth more than $30 billion at current prices. Some companies are off track by more than 50% in certain regions.
The shale boom has lifted U.S. output to an all-time high of 11.5 million barrels a day, shaking up the geopolitical balance by putting U.S. production on par with Saudi Arabia and Russia. The Journal’s findings suggest current production levels may be hard to sustain without greater spending because operators will have to drill more wells to meet growth targets. Yet shale drillers, most of whom have yet to consistently make money, are under pressure to cut spending in the face of a 40% crude-oil price decline since October.
Companies whose wells appear to lag behind forecasts, according to the analysis, include Pioneer Natural Resources Co. and Parsley Energy Inc., two of the biggest oil and gas producers in the Permian Basin of West Texas and New Mexico. The Journal’s review didn’t include some leading producers, such as Exxon Mobil Corp. , because they didn’t make shale-well projections.
Pioneer, Parsley and several other companies disputed the findings, saying the third-party estimates used by the Journal differ from their forecasts on key points such as the likely lifespan of shale wells.
Some companies, including major North Dakota producer Whiting Petroleum Corp. , acknowledged the forecasts can be unreliable and said they were moving away from providing such estimates.
Another North Dakota driller, Oasis Petroleum Inc., said the projections it provided in investor presentations were estimates made as it tested drilling in vast tracts, including areas it has since abandoned. “It’s not a science,” said Richard Robuck, the company’s treasurer. “It’s more of an art.”
Few U.S. shale companies disclose exactly how they make their forecasts—the systems they use and the assumptions they make to estimate well-by-well production—or whether their projections from years ago hit the mark. The fact that many have missed is an open secret in the industry.
“I certainly expect many of today’s estimates will turn out to have been pretty optimistic,” said Francis O’Sullivan, director of research for the MIT Energy Initiative, which has examined shale forecasting. He said the complex geology of shale basins and assumptions based on a small number of wells could make forecasts unreliable. “There is profound variability in the performance of these wells,” he said.
Schlumberger Ltd. , the oil-field-services giant, reported in a research paper that secondary shale wells completed near older, initial wells in West Texas have been as much as 30% less productive than the initial ones. The problem threatens to upend growth projections for America’s hottest oil field, the company said in October.
Oil engineers and reserves specialists say existing data suggests there is a more accurate way to model well output. Operators, they say, must use more conservative assumptions about how quickly production will decline and how many wells can be drilled in a given area. Operators also should avoid making forecasts without a sufficient sample size of wells, they say.
Flawed forecasting doesn’t mean U.S. oil output is about to drop. Shale wells reach peak production quickly and rapidly decline, so companies are constantly drilling new wells. But if thousands of shale wells produce less over their lifetimes, companies will reap less of a long tail than anticipated, requiring them to spend more to sustain output and making it harder for them to reach profitability.
Shale companies have attracted huge amounts of capital from Wall Street over the past decade. So far, investors have largely lost money. Since 2008, an index of U.S. oil and gas companies has fallen 43%, while the S&P 500 index has more than doubled in that time, including dividends. The 29 companies in the Journal’s analysis have spent $112 billion more in cash than they generated from operations in the last 10 years, according to data from FactSet, a financial-information firm.
All oil companies are required to file estimates of total proven oil reserves with the Securities and Exchange Commission. Those estimates, governed by strict rules, generally only capture future reserves companies plan to tap in a five-year period. As the fracking boom intensified, many exploration and production companies looked for a way to persuade investors to value their prospects outside of that five-year window.
Shale companies began touting a metric known as estimated ultimate recovery, or EUR, in investor presentations. The estimates, often represented graphically by what is known as a type curve, project how much oil and gas wells are likely to produce over several decades, including the rate of decline.
The practice of promoting EURs became widespread after oil prices crashed in 2014 and producers, many in need of capital infusions from Wall Street, talked up their prospects. Wall Street’s valuation of many shale companies, which had been closely tied to the value of their proven oil and gas reserves, began diverging.
At the end of 2007, the companies in the Journal’s well analysis that existed at the time had an enterprise value, measured by market capitalization plus debt, of about 1.65 times the value of their proven reserves, according to company disclosures and S&P Global Market Intelligence data. Ten years later, that multiple had risen to more than 2.5 times, even though oil prices were much lower. Last year, the enterprise value of the 29 companies in the Journal’s analysis was $360 billion higher than the value of their proven reserves.
EUR estimates from many companies were grounded on two assumptions: that they could pack wells closer together, squeezing more value from the land they leased, and that they could replicate their best early wells. The results to date suggest those assumptions were often wrong.
The Journal’s analysis involves public data that at times gives an incomplete picture of well performance. North Dakota reports oil and gas production by well, but Texas does so only by land parcel. Third-party data providers must extrapolate to make up for that, meaning their data may not be as precise as well-level data maintained by companies.
The Journal relied primarily on figures from Rystad Energy, but consulted with several other third-party providers, including Oseberg Inc. and BLR Digital LLC, whose data pointed to similar conclusions. Those providers forecast well output over several decades based on early, publicly reported production data, taking into account typical decline rates.
When oil prices plummeted around 75% between 2014 and 2016, to below $30 a barrel, many shale companies used EUR estimates to try to persuade investors that the sector remained a strong place to put their money.
The production forecasts made by many companies were “dangerous” because they were based on a small population of wells, and the performance of individual wells varies significantly, said Norman MacDonald, a natural-resource specialist at asset manager Invesco Ltd.
“Companies were able to high-grade the numbers, show those to Wall Street, and the stock price went up accordingly,” said Mr. MacDonald, a portfolio manager who has urged shale companies to prioritize profits over production growth. “Geology doesn’t line up with Excel spreadsheets too well, unfortunately.”
In September 2015, Pioneer Natural Resources, based in Irving, Texas, told investors that it expected wells in the Eagle Ford shale of South Texas to produce 1.3 million barrels of oil and gas apiece. Those wells now appear to be on a pace to produce about 482,000 barrels, 63% less than forecast, according to the Journal’s analysis.
An average of Pioneer’s 2015 forecasts for wells it had recently fracked in the Midland portion of the Permian Basin suggested they would produce about 960,000 barrels of oil and gas each. Those wells are now on track to produce about 720,000 barrels, according to the Journal’s review, 25% below Pioneer’s projections.
Pioneer disputed the conclusions, noting that it assumes its wells will produce for at least 50 years, while Rystad Energy uses 30 years in its forecasts. Pioneer also assumes its well productivity will fall off at a slower rate than the 7% final decline rate Rystad assumes.
“We find it is simply impossible to compare the numbers due to the methodological differences,” a Pioneer spokesman said.
Adjusting for those factors doesn’t fully make up for the disparity in production forecasts. If Pioneer’s wells produce for 50 years and decline at 5% annually, its current production trajectory would still be 849,000 barrels of oil and gas in the Permian, nearly 12% below the company’s forecast, according to the Journal’s analysis. In the Eagle Ford, estimated production would increase only slightly to 498,000 barrels, or 62% less than the company projected.
A spokesman for Pioneer said problems in the Eagle Ford in 2015 were “widely known,” and the data shows the company’s well performance has improved.
While it is difficult to know how long shale wells will remain productive, assuming tens of thousands of them will pump for 50 years without costly interventions to keep them flowing is extremely optimistic, according to specialists on reserves.
The oldest case study to date is in the Barnett shale in and around Fort Worth, Texas, where modern fracking began about 20 years ago. Researchers at the University of Texas and Rice University predicted that many wells in the region, which primarily contains natural gas, won’t even produce for 25 years. About 73% or more of the total output of wells will come in the first decade, with little value coming after 20 years, the researchers said.
The decline rates of as low as 5% that some shale companies have adopted for their wells are optimistic, some academics and industry leaders say. Recent studies by Rystad and analytics firm Wood Mackenzie Ltd. found that a range of 12% to 16% was the most common decline rate after about five years.
In 2014, Parsley Energy, an Austin, Texas-based producer, told investors its average well in the Midland section of the Permian Basin would produce 690,000 barrels, according to a review of Parsley’s quarterly earnings presentations. By 2015, its estimates averaged 1,050,000 barrels.
Parsley is on track to miss its Midland well forecasts for every year from 2014 to 2017 by an average of 25%, according to the Journal’s analysis.
“Responsible evaluation of the data shows that Parsley’s historical well production has been consistent with expectations set forth in our public materials,” said a Parsley spokeswoman.
When calculating its estimates, Parsley includes other valuable hydrocarbons that come out of wells, such as ethane. In its analysis, Rystad doesn’t include those hydrocarbons, known as natural gas liquids, because they aren’t accurately captured in available public data.
Parsley declined to comment on how much of a boost its estimates get from such liquids. In recent years, the average increase in barrels from including the byproducts amounts to 10% to 18% in the Midland basin, according to third-party estimates.
Mark Papa, a fracking pioneer and chief executive of Centennial Resource Development Inc.in Sugar Land, Texas, said his company avoids forecasts because they create an “illusory picture” of a company’s prospects. “A lot of type curves present a well’s potential under perfect conditions,” he said. “But in reality, the majority of the wells don’t turn out that way.”
Some companies said they were aware of flaws with the forecasting method and how it has been used, but that they provided the numbers to meet demands from analysts and short-term investors such as hedge funds. Some said that if they didn’t, their stock would under perform peers that made optimistic claims.
“You have to make projections,” said David Lancaster, chief financial officer of Matador Resources Co., a top Permian driller. He said companies should revise forecasts that appear to be off target, disclose production ranges rather than specific estimates and avoid screening out poorly performing wells.
Matador’s average well in the Permian’s Delaware Basin is on track to outperform forecasts in all three years the company provided them, according to the Journal’s analysis.
Denver-based Whiting Petroleum is de-emphasizing its production estimates at the direction of its chief executive, Bradley Holly. Mr. Holly, who became CEO in November 2017, said the company is now more focused on generating cash, lowering debt and maximizing a well’s returns early in its life.
“Your return will really be made in the first two to three years,” he said.
One reason thousands of early shale wells aren’t meeting expectations is that many companies extrapolated how much they would produce from small clusters of prolific initial wells, according to reserves specialists. Some also excluded their worst-performing wells from the calculations, which is akin to eliminating strikeouts when projecting a baseball player’s batting average.
“There are a number of practices that are almost inevitably going to lead to overestimates,” said Texas A&M University professor John Lee, an expert on calculating oil and gas reserves.
Many reserves specialists have advised companies to provide a potential range of outcomes based on their internal analyses, a common practice in statistics that better accounts for uncertainty.
Academic research has suggested that data from at least 60 wells, producing for six months or more, would be needed for accurate forecasts. Yet some companies and analysts have made predictions based on fewer than 10 wells.
At a July presentation Mr. Lee gave in Houston about techniques that could produce more accurate shale forecasts, one participant stood up and challenged the engineers in attendance.
“Why aren’t we doing this?” the man asked several times, according to Mr. Lee and two other people who attended the meeting.
“Because we own stock,” replied another engineer, sparking laughter.
Corrections & Amplifications Enterprise value of selected U.S. oil companies in 2017 on average was 2.8 times the value of proven reserves, compared with 1.7 times in 2007. A graphic with an earlier version of this article incorrectly stated the values were 2.8 times larger than reserves and 1.7 times larger than reserves, respectively. Additionally, an average of Pioneer Natural Resources Co.’s 2015 forecasts suggested company wells in the Midland portion of the Permian Basin would produce about 960,000 barrels of oil and gas each. An earlier version of this article incorrectly stated on second reference that the company’s production forecast for wells in the region was 849,000 barrels. (Jan. 7, 2019)