A recent posting from the Strong Towns organization pointed out that we, as a society, criticize public transit as being expensive for very little utilization, but that, in the author's opinion, roads are a worse investment based on passenger mile costs. But, in my EZ opinion, passenger miles is a dangerously incomplete tool with which to measure the value and utility of any transit system.
What I would like to see is a study that would look at usage/riders/freight in comparison to surrounding population density. For instance, in Texas the State Farm to Market Highway System may be very, very lightly used, but those roads made it possible to move agricultural goods efficiently to market, feeding a large part of the city populations of the world and providing great economic benefit to the farmers and ranchers. I don't think it makes sense to look at whether or not the system is "empty" without looking at other factors including total economic cost/benefit and surrounding population density. Further, to say that a road benefits ONLY those who choose to live along or at the end of it is specious, shallow thinking. Our economic system is much more complex than that. You've seen the signs that read, "If you have eaten today, thank a farmer." Yes, but also thank the road builder that made it possible to move that food to your store and home and the taxpayers who pay for it. As a side note, I think that an appropriate fuel or per-mile tax is the right way to pay for roadway infrastructure - any increased costs for commercial traffic would simply be passed along to the consumer, without whom the goods wouldn't be moved in the first place. Thus, traffic on very high usage roads would "subsidize" the costs of lightly used, but important, roads. From those funds collected, we then decide where to best spend them. When we consider these things, please consider that passenger transit is a side blessing -- the main purpose of roadways is economic and military.