The reality is that many benefits -- like increased security, reduced risk of accidents and improved asset utilization -- are often misunderstood, misstated, and given little weight when a company makes a decision about which technology to deploy. Instead of viewing them as "soft" factors with little more than public relations value, companies would do well to include such seemingly intangible costs when analyzing the benefits of installing a new system.

Witness the case of Burlington Northern Santa Fe (BNSF), the freight railroad that pioneered automated means of tracking and controlling the passage of trains in the late 1980s and early 1990s, then opted not to deploy it. Depending on who you talk to, the reasons for rejecting deployment were economic, technical, political and cultural.

And possibly fatal. Just last month, a BNSF freight train careered into a southern California commuter train, killing two people and injuring 100. In early May, the family of a passenger who suffered a brain hemorrhage in the course of the accident filed suit seeking compensatory and punitive damages from BNSF, alleging the railroad consciously chose not to install automated braking technology that could have prevented the accident because of the costs.

BNSF now may want to reevaluate its decision against implementing automation that could control trains already en route. The railroad declined to comment on its plans other than to say, through spokesman Pat Hiatte, that it is continuously reviewing its technology deployment strategies.

Yet the question remains: What is a reasonable cost for the kind of technology that can prevent passenger fatalities?

A single lawsuit may not sway BNSF toward adopting so-called positive train control (PTC) technology. After all, when the railroad nixed such a system nearly two decades ago, it seemed like a profitable idea. Economic and feasibility analyses found PTC would have helped the railroad achieve a 30% return on its investment after only three or four years.

PTC systems are designed to trace and manage passenger and freight trains by using sensors, computers and digital communications to control trains, braking systems, grade crossings and planning and scheduling systems. These control systems are the hub of what the Federal Railroad Administration (FRA) calls intelligent railroads. Among other functions, PTC systems allow dispatchers to override train crews in the event of an emergency-applying brakes automatically, for instance, if trains were on a collision course. As in the California crash.

Such systems would be costly. The FRA, which is a safety branch of the federal Department of Transportation, estimates it would cost railroads $3 billion to $5 billion over a decade to implement PTC systems throughout the U.S.

In 1982, Burlington Northern undertook a project to test the feasibility of a PTC system using defense satellites to track trains. Burlington Northern installed its Advanced Railroad Electronics System (ARES) prototype on its Iron Range route through Minnesota and operated it from 1987 to 1989. The railroad spent approximately $15 million on ARES, while its technology partner, Rockwell International, had spent three times that amount, according to a Harvard Business School case study of the project. As a result, Burlington Northern estimated that deploying ARES nationwide would cost $350 million.

Burlington Northern at the same time installed a state-of-the-art network operations center in Fort Worth, Texas. That 180,000-square-foot, disaster-proof center still exists, and operates as the nerve center of BNSF's dispatching operations.

Burlington Northern brought in a battery of consultants to help the railroad perform a variety of analyses to determine if fully deploying ARES made sense. The consultants estimated everything from the effect of ARES on the efficiency of carrying freight, to impacts on labor, safety and capacity.

The analyses, detailed in the Harvard study, indicated "the potential benefit of ARES is large but uncertain." In the end, the gross benefit was pegged at $400 million to $900 million over the life of the system, with an expected present value of about $600 million. Those benefits greatly outweighed costs that were estimated to represent $220 million in upfront or "present value" dollars, at time of deployment.

But Burlington Northern's top brass wasn't convinced by the findings, according to the Harvard study and interviews with some of those connected with the ARES study. The case study researchers quote Dick Lewis, then vice president of strategic planning, as calling ARES "a technology in search of a problem." Then-Chief Operating Officer Bill Greenwood said ARES technology was "the least of our concerns," and expressed dismay that implementing ARES would require Burlington Northern to change its underlying business processes.

"Railroads heavily prefer hard-dollar savings to soft-dollar savings," says Steven Ditmeyer, head of the Federal Railroad Administration's Office of Research and Development-and former chief engineer of research, communications and control systems for Burlington Northern. "That was true then, and it's still true now." Ditmeyer identifies soft-dollar savings as improved running times, greater reliability, improved service, fewer train collisions and the like. For railroads, hard-dollar savings are monies saved from, for example, layoffs and reduced fuel costs, he says.

Ditmeyer and the FRA are still pushing for railroads to adopt PTC and similar intelligent railroad systems. The FRA is funding two PTC demonstration projects at the moment: one in Michigan, in conjunction with Amtrak, called the Incremental Train Control System, and another in Illinois, jointly with Amtrak and the Association of American Railroads.

Outside the government, the debate over whether PTC systems are cost-justifiable continues to rage, just as it has for years.

"Railroads have been making arguments about why they can't or won't implement these kinds of (PTC) systems since Burlington Northern did several analyses nearly two decades ago," says one transportation analyst, who requested anonymity. "But there is no economic reason why they haven't done this. These systems would pay for themselves in three to four years."

In the end, however, improving safety-one of the main benefits accrued from PTC systems-doesn't fare well in cost/benefit analyses.

Carl Martland, senior research associate at the Massachu-setts Institute of Technology, says the FRA has estimated railroads could save an estimated $20 million to $50 million a year, based on a $2.5 million per fatality cost figure. Even there, passenger lines are the ones likely to reap the greatest savings from improved safety. According to preliminary 2001 FRA data, there were only 2,962 accidents involving trains, with six resulting fatalities (not including accidents and deaths resulting from highway-train and/or crossing-grade incidents).