A major rationale — perhaps the major rationale — touted by supporters of mass transit is that by reducing our output of greenhouse gases and other pollutants, transit can help save the environment. The proposition seems intuitive and even obvious: by no longer encasing each traveler in thousands of pounds of difficult-to-move metal, surely transit is more energy-efficient. Plenty of analyses prove this. But then again, Aristotle, who was revered as the infallible font of truth for more than 1,000 years, proved that heavier objects fall faster than lighter ones and that women have fewer teeth than men. Might studies that demonstrate transit is greener be similarly wrong?
 
They might. The reason is that many studies of energy efficiency by mode often make questionable and — depending on the author’s point of view — self-serving assumptions. The main trick is to look at autos with but one passenger and compare them to transit vehicles in which every seat is full. (For example, see this.)
 
But in the real world, this is emphatically not the case. At any given time, the average auto has somewhere around 1.6 passengers, and the average (typically 40-seat) bus has only about 10. Rail vehicles typically have more passengers (on average about 25), but then again they are also typically much larger. Thus their average load factor (percentage of seats filled) is also not high, at about 46 percent for heavy rail systems (think subways in major cities) and about 24 percent for light rail (think systems that mostly run at street level).
 
It is not clear that moving around large and largely empty vehicles is much of an improvement over moving around smaller ones. In fact, it may be worse. According to the Department of Energy’s Transportation Energy Data Book, in 2010 transporting each passenger one mile by car required 3447 BTUs of energy. Transporting each passenger a mile by bus required 4118 BTUs, surprisingly making bus transit less green by this metric. Rail transit admittedly fares better, at 2520 BTUs per passenger mile, but even this is not the kind of slam-dunk advantage over the auto that transit advocates might hope for.  
 
There are some qualifications here, some of which aid transit and some the car. The sources of transit’s energy—electricity in the case of rail, and often natural gas in the case of bus—are typically cleaner than those of the auto, so that even with higher energy usage, transit may produce fewer greenhouse gases. Still, even taking this into account, a bus produces more CO2 per passenger mile than the car for most trips.
 
Transit looks better when the environmental costs of building the vehicles and infrastructure are spread over the passenger miles that each mode will accommodate. Taking this into account, an average bus at normal load factors does currently produce slightly fewer pounds of CO2 than autos, and rail transit does much better.
 
On the other hand, electricity is not much cleaner than gasoline if it is derived primarily from coal, and in some states, such as—surprise—West Virginia, nearly all of it is. And in any event, the electrification of autos is advancing. Moreover, greatly increased fuel economy, even for regular old internal combustion vehicles, is rapidly making cars more efficient. By the time the new fuel economy standards for autos are reached in 2025, a new car’s average mpg will have increased from 27 to about 40, meaning that in terms of energy efficiency, cars may leave buses behind and begin to be competitive with rail.  
 
Does this mean that efforts to increase ridership will not help—and may even harm—the environment? Not so fast. In fact, the environmental benefits of higher transit ridership might be very great—but they depend completely on how those increases are obtained.
 
Pumping up ridership by adding transit service will probably do little good, and may even be counterproductive. The reason is that new service promises to reduce transit’s already less-than-spectacular load factors and result in largely empty vehicles.
 
Why? At this point we have picked all the low-hanging fruit in terms of transit markets. We already have extensive rail transit service in the places where land use and demographic characteristics are congruent with high transit use—such as central Chicago, Boston, the San Francisco Bay Area, Washington D.C., Philadelphia, Los Angeles, and (writ very large) New York City. For the most part, any new transit service has to go to relatively low-density cities and low-density areas within cities, meaning that new investment would drag transit’s overall efficiency down, not up.  
 
To give an idea of how this phenomenon works, the heavily used New York subway system (58 percent of seats are typically filled) produces .171 pounds of CO2 per passenger mile, less than 1/3 the average for cars nationwide. However, the much more lightly used Cleveland, Pittsburgh, and Memphis light rail systems actually produce considerably more CO2 per passenger mile than cars do. Unfortunately, due to diminishing returns, new rail transit systems are more likely to resemble the latter three than the former one.
 
On the other hand, if we can persuade travelers to leave their cars and ride existing transit service, rather than new service, the environment will benefit greatly. Given its current low load factors, transit generally has plenty of capacity to absorb new customers with practically zero additional energy expenditure.
 
Strategies to pursue this would involve economic incentives to influence behavior. These might include pull strategies designed to lure riders onto transit, such as fare cuts, or push strategies designed to get them out of cars onto transit, such as increasing gas taxes, congestion tolling, or charging market rates for street parking. Note that pricing strategies of this sort would also help the environment by raising auto efficiency through increased incentives for carpooling.  
 
Thus it makes more sense to focus on policies that attempt to change travel behavior as opposed to building infrastructure and buying vehicles. However, public policy in recent decades has self-defeatingly focused on the latter strategy, not the former. This has proven very popular, since shiny new trains and buses allow politicians to take credit for very visible improvements, while the costs of said infrastructure and equipment are essentially invisible since they are spread broadly across the American public and are hidden in arcane budgeting processes. (With a thicket of passenger fares, fuel taxes, sales taxes, bonding, advertising revenue, etc. coming from federal, state, regional, county, transit district, and municipal administrations, it can be difficult even for experts to determine who pays what for transportation
.)
 
So there it is: to benefit the environment, probably the best thing to do is be very skeptical about adding new transit service and even to discontinue some service we are currently providing (sorry, liberals). Simultaneously, we should raise fees and taxes for driving (apologies to you conservatives). The best I can offer to keep the comments section below free of hate posts is that at the moment the chance of either happening is small. This may make good sense given the current political atmosphere—but it makes little sense for the atmosphere of the planet.