I am (currently!) mildly against Keystone XL. Why? Simply put: American railroads have proven themselves up to the task of moving lots of crude oil at a remarkably low cost. The oil will flow regardless of the pipeline.
Now, there is a debate over whether railroads can move the oil from the Canadian tar sands. I will leave that for now. This post is about the United States. After all, moving Dakotan oil is often cited as a reason to build the pipeline.
In the U.S., trucks and railroads have already done a much better-than-expected job of moving crude around. Here is a chart of refinery loadings by transport mode:
Of course, non-pipeline deliveries only accounted for 7% of total deliveries last year. A skeptical reader might argue that they are still a sideshow. And they might be right, were it not for the price data. Consider first that the spread between Brent (a proxy for the world price of oil) and WTI (a proxy for the American price of oil) has collapsed:
The implication is that crude is no longer bottled up in the United States. Of course, an even more skepital reader could argue that crude is still be backing up in North Dakota: the Eagle Ford in Texas is what has pushed WTI back up to world levels. That is possible, except for this:
The differential between the price in North Dakota and WTI essentially disappeared sometime in 2012. (Note for pedants: Before the big production explosion, Bakken oil was more expensive than WTI. That changed when production took off and strained the transport infrastructure.)
In short, trucks and trains have done a pretty good job of integrating the new North American energy geography. Who needs Keystone? The North Dakotan oil is flowing and the Canadian oil will flow.
Well, one remaining argument for the pipeline is that it is safer. (We will ignore trucks for the moment: the general agreement is that more train transport is what will pick up the slack in a no-Keystone world.) There is some evidence for the proposition. Consider:
That is a huge increase! Thing is, the amount of oil moved has also seen a huge increase. Moreover, most of the incidents do not involve actual spills. (None involve fatalities.)
When you do the math on spillage (by volume transported) you get 0.0005% for pipelines and 0.0009% for railroads. Thing is, most of the railroad spills come from one single 2008 spill in Oklahoma. Take that out, and you have 0.0001%. Here is an article that discussed the safety of the two transport modes; it could win an award for Most Misleading Headline.
Now, these are per-gallon, not gallon-mile. It is hard to compare the spill figures, because pipelines do not need to report spills below five gallons unless they involve injuries or more than $50,000 in damage. With that caveat in mind, we can compare rates for all crude spills. There we get over 2002-12 a rate of 0.88 gallons per billion barrel-miles for pipelines and 0.38 for railroads. (Of the 129 railroad incidents, 35 were over 5 gallons. Of the 1,849 pipeline incidents, 1784 were over 5 gallons ... but remember the caveat that the under-5 pipeline reports had to involve injuries or more than $50,000 in damage. The link is to an industry source but the underlying data come from the federal government.)
What about the recent Quebec disaster? Well, the tanker cars used up there have long been known to be dangerous. The problem is easy to fix. (It is unforgiveable that the disaster was allowed to happen.) It is not a reason to prefer pipelines.
Given the available data, trains do not seem to be more dangerous than pipelines. (To be frank, I would not jump to conclude from the above data that they are safer; just that there is likely no real difference.) Moreover, there are some reports that Transcanada might be cutting corners on Keystone.
On balance, therefore, I am currently against the project ... but I have changed my opinion before, and new evidence could change my mind again. If there is demand, I will examine whether the presence of all that oil in Alberta alters or reinforces my current opinion.