At first glance, high-speed trains between Phoenix and Tucson sound like a great idea. I love trains, and the idea of boarding a comfortable train that will whisk me to my destination seems very attractive.

But I am also an economist, so I have to ask, “Who will pay the cost? Who benefits? And do the benefits justify the costs?”

There is a good reason autos and airplanes have largely replaced passenger trains. Rail lines, especially high-speed lines, are extraordinarily expensive to build, maintain and operate. Justifying these costs requires huge volumes of traffic, and our decentralized world has few places dense enough to support such passenger traffic.

Take the Boston-to-Washington corridor, by far the densest in the nation. Amtrak’s Acela may be the fastest way to get from downtown to downtown, but few people live or work downtown anymore, so the Acela carries only about 2 percent of passenger traffic in the corridor.

To cover its operating costs (but not its capital costs), Acela fares from New York to Washington begin at $139. By comparison, JetBlue fares begin at $39, while a variety of bus companies offering rides for $15 to $20 carry almost 50 percent more passenger miles than Amtrak. Buses take about 80 minutes longer than the Acela but offer free wireless Internet so travelers’ time isn’t wasted.

Amtrak carried the equivalent of more than 10 million trips in the Boston-Washington corridor in 2010. The Arizona Department of Transportation estimates a Phoenix-Tucson high-speed train would attract, at most, 1.9 million trips a year, or about 2,600 round trips per day, in 2050. That is nowhere near enough to cover costs, so the trains would require millions in annual operating subsidies as well as hundreds of millions, if not billions, in capital subsidies.

Nor are trains particularly environmentally friendly. Intercity buses use 60 percent less energy per passenger mile as Amtrak trains, and when full life-cycle costs are counted, the difference is even greater. Autos are getting more energy-efficient each year, and by 2025, the average car on the road will use less energy per passenger mile than any high-speed train.

Tourists riding the high-speed trains in France and Japan have come home with fantasies of American high-speed rail. Yet high-speed trains don’t work very well in those countries, either. Most require significant subsidies, and the average residents of France and Japan ride the TGVs or bullet trains less than 500 miles, equal to about one round trip, per year.

I recently visited Japan and rode both the subsidized bullet trains and the parallel conventional trains that, because Japan is so densely populated, operate at a profit. Most riders of the conventional trains were students, senior citizens and working-class families, while most riders on the high-speed trains were well-to-do business travelers and foreign tourists.

For the vast majority of Arizona residents, building high-speed rail means taking their tax dollars to subsidize trains they won’t be able to afford or find convenient to ride. Those who do ride the trains will be relatively wealthy people who don’t need your subsidies. As much as I personally love trains, I have to conclude that high-speed rail is a bad idea for Arizona.