It has always seemed obvious that substantial energy savings could be realized by diverting commuters from cars onto rail transit. In fact, the wisdom of this idea has appeared so self-evident, to so many people, that is has been little examined. The only direct analysis (1), calculated the energy-cost of various kinds of transit construction and concluded that the United States could save energy by diverting investment from highways to rail transit.
Those conclusions were based on three implicit assumptions, but a reasonable modification of those assumptions produces a directly opposite conclusion. Their unstated assumptions were: 1. that Congress often acts as if the expenditure of construction money were an end in itself (while I will assume that the expenditure should be evaluated in terms of the passenger-services it produces); 2. that engineering capacity measures are appropriate for estimating patronage (while I will use observed behavioral data because there is little evidence to support the continued hope that the public's demonstrated dislike of "public transit" can be altered to any significant degree (2)); 3. that modern rail systems would have similar energy characteristics to existing rail systems (but data which became available after their article make it clear that modern systems are actually much less efficient (3)).
My analysis uses data from the San Francisco Bay Area Rapid Transit system (BART), and evidence is presented to show that BART is typical of modern rail systems. I shall first concentrate on the construction energy relationship and will show that for a standarized measure of services, passenger-miles, freeway construction is much more energy efficient than rail transit construction. I will then show that viewed as a total system--looking at the energy involved in propulsion, building automobiles and other transit vehicles, and right-of-way construction--BART should never have been built in the first place if energy saving was the only goal.