I’ve written extensively on this blog about some of the ridiculous new arguments being peddled by opponents of the Keystone XL oil pipeline. I hadn’t seen as much in the way of new and implausible assertions from pipeline proponents – until now. Here’s what I found in my email yesterday courtesy of the U.S. Chamber of Commerce:
“The project [that environmentalists] oppose is construction of the Keystone XL Pipeline, which would carry crude oil from Canada to refineries primarily along the Gulf Coast and immediately create 20,000 jobs along the pipeline route. Furthermore, economic impact studies show that 250,000 permanent jobs will be created over the long term.”
The “economic impact study” in question appears to be a widely cited report by The Perryman Group. (A second widely cited report, by the Canadian Energy Research Institute (CERI), gives a figure of 270,000 jobs but is not directly germane to Keystone XL – it projects impacts from a series of pipelines that might be built between now and 2030.) The Perryman report has been criticized for the claim of 20,000 jobs along the pipeline route. I’ve seen less criticism of the far more impressive 250,000 number (though this is one exception). That’s a shame, since while the number is being invoked prominently, the analysis upon which it’s based is dead wrong.
Critiquing the Perryman study is a bit tricky since its methodology is opaque. When I run into this sort of problem, I make sure that I can reproduce the reported results, before I go on to audit them. The Perryman report describes its methodology thusly:
“The Keystone XL Project would lead to positive outcomes for US energy markets by providing access to a stable source of incremental petroleum supply, reduced risk and, thus, price. Furthermore, operations of the Keystone XL pipeline will provide a significant ongoing economic stimulus including the creation of hundreds of thousands of jobs due to the stable oil supplies it will make available.In order to model these effects, The Perryman Group initially examined the likely effects of such a shift by calculating the magnitude of the increase in availability (based on an 80% capacity factor for purposes of conservatism) relative to anticipated domestic consumption in the 2010-2012 timeframe. The price effect was then derived using elasticity coefficients for the US market. Because this supplemental supply represents a permanent change in the market, a long-term response parameter was adopted. This measure was obtained from academic research and independently verified for reasonableness.”
The Keystone XL pipeline would add 700,000 barrels per day of transborder pipeline capacity. With an 80 percent capacity factor this becomes 560,000 barrels per day. The Perryman report asserts that U.S. consumption is 18.7 million barrels per day (though in one place it says that the number is 20.7). Perryman assumes a baseline oil price of $66.52 per barrel, and the academic research referenced puts U.S. oil demand elasticity at -0.453. Mix this all together and you get a projected price decline of $4.40. Multiply this by 18.7 million barrels per day and, presto, you get an annual impact of $30 billion, pretty much the same number that Perryman uses to ultimately generate 250,000 jobs.
There are, however, several things wrong with this calculation. For starters, it assumes that Canadian production will rise by the volume transported by Keystone XL if and only if the pipeline is built; in reality, over time, alternative pipelines are likely to be built, blunting that impact. On a more technical but more fundamental front, U.S. consumption and demand elasticity are the wrong numbers to use in determining the price impact of new supply. The United States is part of a global market, so Perryman should be using global, not U.S., figures. The relevant numbers are about 90 million barrels per day of global demand and a global demand elasticity of somewhere around -0.4. This leads to a projected price impact of about one dollar. But that’s still too high: in response to lower prices created by more Canadian production, both OPEC and non-OPEC (ex-Canada) production should drop. Let’s say that this removes half the impact of new Canadian production on world supply. The price impact of the new Canadian oil supply is now fifty cents a barrel – barely 10 percent of what Perryman projects. Given the Perryman methodology, this should cut the jobs impact by a similar factor, to somewhere around 25,000 rather than the 250,000 originally projected. It’s also worth noting that somewhere between 25 and 50 percent of the material supplied by Keystone will be diluent rather than Canadian oil, which should decrease the impact even further, to somewhere between 12,000 and 19,000 jobs.
One might also contend, of course, that the demand elasticity is higher or that the supply elasticity is lower than I assume; more on that in a moment.
First, let’s take a look at several other problems with the Perryman analysis. Perryman asserts that savings by U.S. consumers on oil end up getting injected into the U.S. economy. Over the long term, though, that should be partly offset by losses to U.S. oil producers, who must also absorb the lower prices. U.S. liquids production is equal to about half of U.S. consumption; this means that we need to eliminate another 50 percent of Perryman’s projected jobs. We’re now down to somewhere around 7,000 or 8,000 of them. Of course, the money that the United States sends abroad for oil doesn’t vanish – some of it comes back to buy U.S. goods and services. That phenomenon should erode the jobs impact of Keystone even further.
The last big problem with the Perryman analysis is its conversion of the oil savings figure into macroeconomic impacts. This one is more difficult to pin down, since the model isn’t available, but the final figures that Perryman reports are mighty suspicious. For starters, he projects that lower oil prices will boost U.S. oil and gas production profits by more than six billion dollars. This makes no sense: lower oil prices should hurt, not help, oil producers. He also reports that total spending in the U.S. economy rises by $100 billion as a result of lower oil prices even though total GDP rises by only $29 billion. The net impact of that would need to be a $71 billion increase in the U.S. current account deficit (whatever you spend money on that you don’t produce is something that you have to import). How lower oil prices can increase the current account deficit is beyond me. It suggests that something in the model is awry.
I mentioned earlier that one might challenge my elasticity figures. Let’s say, very pessimistically, that global demand elasticity is -0.2, and that non-Canada supply elasticity is zero. We still get a jobs impact of only 40,000, and that’s setting aside the impact of fixing the broader macroeconomic model. Once again, the 250,000 figure is far too high.
Let me be clear: I’m not claiming that Keystone XL will create 7,000 or 8,000 or 40,000 jobs. I find the entire approach of the Perryman study suspicious. What I’m saying is that even if you buy its overall methodology, fixing the basic numbers leads you to much lower jobs estimates. In reality, the actual jobs impacts might be lower or higher than the numbers I’ve shown here. What’s for certain is that the 250,000 figure isn’t supported by the analysis presented.
And I ought to reiterate another basic point. I don’t see any reason to block the Keystone XL pipeline, so long as local concerns in Nebraska are fairly addressed, something that shouldn’t pose a high hurdle. The Keystone XL debate is a distraction from things that really matter to the future of U.S. energy and climate policy. So long as the debate is front and center, though, correct facts would be nice.