Eric Morris is only the most recent wannabe transit expert to discover that empty buses aren't very efficient. His analysis is strikingly reminiscent of Tom Rubin's, but he brings to the topic a splash of condescending Freakonomics overconfidence, and that special ignorance of real-world transit that only someone who's driven to Brian Taylor seminars can pull off.
Back in 2009, Kevin Libin used Rubin's numbers (with proper credit) for a National Post transit hit job. It was wrong then, and it's wrong now.
Of course, Morris and Rubin are right that it's dishonest to compare the efficiency of a single-occupant car with that of a full capacity bus. But the report that they criticize (PDF, page 4) is careful to give figures for buses with one, five, eleven, forty and seventy passengers, and in the Reason "debate" the authors insist that they were not making such a comparison.
Morris and Rubin quote figures from the National Transit Database that the average bus in the United States has ten people on it (about forty people can sit on a bus, and seventy can fit on if some people stand), for 25% occupancy, and light rail has 24%. Morris acknowledges that heavy rail (metro) systems have 46% occupancy, but he doesn't mention that rail systems are much more energy-efficient than buses.
25% occupancy is also not normal for a bus outside of this country. In places where transit is more widely used, there are more than ten people on average. For example, in Zurich and other Swiss cities (PDF) there are 14 on average (35%). Trams carry 53 people on average, and if you assume that each tram can seat 100 people, that's 53% occupancy. In the Czech Republic and in major African cities (PDF, bus occupancy ranges from 63-80%.
There are actually quite a few systems right here in the US that have higher average occupancy rates than 25%. In fact, there are twenty with rates over 40%, including Morris's own LA County MTA and Brownsville, Texas. Here's a graph of the occupancy from 2007 (given in passengers, not percentages, so you have to divide by 40) with the farebox recovery ratio:
So Morris has three straw men here: the report that compares a single-occupant car to a fully loaded bus, picking on the inefficient bus, and using US bus occupancy figures. But he actually admits that 25% occupancy is not destiny, even in the US, in the second half of his post: "Given its current low load factors, transit generally has plenty of capacity to absorb new customers with practically zero additional energy expenditure."
There's a lot more to say about this post. The funniest thing is that Morris's general thrust is actually the main point that I've been hammering at for years: you can accomplish more of your transit goals by restricting or pricing car use than you can by building more transit. He's right: I hate his post. Not because I disagree with his point, but because he does such a sloppy, condescending job arguing it.