More on Counting Shale Jobs and Deciphering Who Cooks the Books

Stephen Herzenberg |

As the saying goes, “It’s better to keep your mouth shut and appear stupid than open it and remove all doubt.”

The last time I recall talking to Dennis Roddy was when he was still a member of the Corbett communications team and phoned the Keystone Research Center (KRC) office. At the time KRC was engaging on the opposite side of a debate over public sector pensions with the Corbett administration. Mr. Roddy let on that he knew our eight-person non-profit provides its employees with a 401(k)-style pension plan. When I stopped laughing, it occurred to me that this was an attempt to intimidate us.

Call me old-fashioned but I hang onto the idea that we’re better off when policy and political debates are based more on logic, evidence, and transparency, and less on playing fast and loose with the facts, attempts at intimidation, or personal attacks.

In Mr. Roddy’s latest column on shale jobs – which we recommend you read and then read again – he wants you to believe that the Wolf Administration has politicized the counting of shale jobs and that the Corbett Administration was as pure as the driven snow when it came to letting the numbers people in the Pennsylvania Department of Labor and Industry (PDL&I) just provide “the facts.” Even if you don’t know what happened, Roddy’s story has the ring of fantasy about it.

But I do know what happened.

At the end of the Rendell Administration and into the Corbett Administration, there was a great hunger for information on the shale boom. In response to this customer demand, the entrepreneurial civil servants leading the data shop within PDL&I, the Center for Workforce Information and Analysis (CWIA), put together a new report called Marcellus Shale Fast Facts.

This booklet included numbers on employment in oil and gas extraction and pipeline construction, six detailed industries that CWIA labelled the “Marcellus Core.” Everyone – CWIA under Rendell, Corbett and Wolf, KRC, academic experts, even Dennis Roddy, accepts employment in these six “core” industries as a good estimate of direct jobs in shale extraction itself plus in pipeline construction.

To estimate Marcellus Core employment requires two steps. First, add up total jobs in the six industries; second, subtract jobs in those industries as of 2005 or 2006 before fracking began in earnest, since those jobs were in conventional gas extraction. (Since separate industry categories do not exist for conventional and unconventional gas extraction, everyone has to assume that all employment change in these six codes since about 2005 is due to fracking not conventional oil and gas extraction.)

As well as “Marcellus Core” employment, however, Marcellus Shale Fast Facts also included lots of other numbers – in fact, pretty much the kitchen sink of numbers related to shale. While heavy on numbers, the booklet was light on interpretation, leaving reporters, industry and others to try to make sense of the numbers largely on their own.

The first episode in misinterpreting Marcellus Shale Fast Facts numbers, ignoring the difference between “new hires” and “new jobs,” is what drew KRC into the shale jobs debate. We published this June 2011 brief correcting the error — which earned us this personal attack from the Marcellus Shale Coalition and another one from the state Republican Party, both of which we responded to here. Who knew that correcting the factual record was a contact sport? (For more on this particular episode, read this op ed, which is a companion piece to this blog.)

The more significant misinterpretation of Marcellus Shale Fast Facts, which Roddy goes to tortured lengths to defend, is an estimate of employment in Marcellus shale “ancillary industries.” As defined by CWIA, the 29 Marcellus Shale ancillary industries are all industries in which shale extraction creates some supply chain jobs – industries such “engineering services,” “trucking,” and “highway, street and bridge construction.” While shale does create some jobs in these and other ancillary industries, most jobs in these industries have nothing to with shale – e.g., UPS drivers are part of trucking but not part of the shale jobs footprint.

Piling absurdity on top of absurdity, to get numbers over 200,000 you also have to count all jobs in these industries, not just the increase since fracking began around 2005. So apparently fracking created about 170,000 ancillary and core jobs even before unconventional drilling began.

Using this method to estimate shale’s jobs footprint isn’t an exaggeration it’s just making stuff up. Yet as recently as January 22, this employment number was used to support a Marcellus Shale Coalition claim that the shale industry supported 243,000 jobs in 2014.

In stopping the routine reporting of jobs in Marcellus Shale ancillary industries, the Wolf Administration has simply stopped publishing numbers that don’t mean anything and have proved susceptible to misinterpretation. This positive, common-sense step also means that CWIA will no longer unwittingly echo invented and exaggerated shale jobs claims.

Dennis “Through the Looking Glass” Roddy, sees the story differently.

First, he misrepresents Marcellus Shale Fast Facts as a “scientific” document produced by “statistical wizards.” He then goes into a long “how I learned the secrets of the temple” explanation of the standard economists’ method of estimating total jobs impact.

While it may be mysterious to Dennis, it is well known among economists that the total jobs impact includes three components: direct jobs, supply chain or “indirect” jobs, and the “induced” jobs created in consumer industries when workers, owners, and royalty recipients in the shale industry and its suppliers spend their additional income. Roddy ends his discussion of this standard economic method with the sentence “that was the context of those numbers” and then quotes Patrick Henderson, Corbett’s energy executive at the time and now with the Marcellus Shale Coalition, “We have about 240,000 Pennsylvanians working in industries that are supported directly, or made more secure, by the growth of oil and gas activity in Pennsylvania.” Roddy seems to be trying to suggest that the 240,000 figure comes out of using the economists’ standard methodology for measuring total jobs impact. But that is not true.

In addition, Roddy doesn’t appear to realize – or at least acknowledge – that the Wolf Administration has actually just done what his opinion piece suggests should be done. The new Wolf methodology DOES explicitly estimate indirect and induced jobs and then add those to direct shale jobs to get the total jobs footprint. The current estimate for direct jobs is 22,058 jobs (down a bit from a year earlier), with an additional 15,706 jobs at suppliers and 19,912 “induced” or consumption-based jobs. That’s a total of 57,676 jobs. In sum, the Wolf Administration is not underestimating the total jobs impact of shale. It has just set aside a ridiculous approach that led to estimates more than four times a legitimate estimate.

Even if Dennis Roddy is living in his own fantasy world, other sources have recognized that the change in methods makes sense. For example, the Times-Tribune calls the new approach a “more honest reporting method.” Several economists have also noted that the new figures create a “much more accurate” employment picture of the shale industry.

At the end of the day, going “inside (the mind of) Dennis Roddy” doesn’t matter. What does matter is the combination of factual distortion and attempts at intimidation that are aimed at blocking lawmakers from enacting state natural gas policies that benefit Pennsylvanians rather than simply cater to the industry. We are still a good ways from achieving the right outcome on that front.