While we do not yet know the precise cause of the recent Amtrak derailment in Montana, the response was swift. The National Transportation Safety Board (NTSB) immediately dispatched a team to investigate the incident which should take about a month. Coverage of the three passenger deaths is robust and spotlights an ongoing concern about railroad safety in the United States. But this type of wreck, while certainty tragic, is quite rare and most fatalities on the rails are the result of other causes.

About three-quarters of the track mileage on which Amtrak trains run is owned by someone else, mostly private freight railroads. These companies’ safety-related investments, upgrades, and modernization efforts over the past few decades, coupled with strong federal standards, means the United States railway network is one the safest modes of transportation for workers, riders, and the public. Compared to trucking, railroads are about four times safer on a fatality basis.

Still, an average of 16 people die somewhere on the nation’s 140,000-mile rail network weekly, and more than four derailments—mostly at low speeds in freight rail yards—occur every day. These incidents are all reported to federal and state authorities but rarely make the evening news. This is because most railroad safety events are isolated incidents that directly affect relatively few people.

Data from the Federal Railroad Administration (FRA) shows that incidents on all U.S. railroads declined nearly 60 percent over the past 30 years. In the last decade, the total number levelled off at about 12,000 annually, at a time when rail traffic increased. Unfortunately, while overall incidents have dropped, the number of fatalities has not. Railroad incidents that result in the loss of human rose 31 percent since 2012.

But less than one percent of railroad incidents that resulted in death in 2019 occurred because of a train accident, like the one Montana. The largest share of fatalities comes from trespassing on railroad property, a problem that has gotten significantly worse over the past ten years and in 2019 killed 549 people. Trespassers enter railroad property for myriad reasons, including because it may be the most direct route to their destination on foot or to loiter (which is often unreported). Sadly, about 30 percent of these fatalities are individuals who intentionally took their own lives. According to the FRA, railroad trespassing collisions cost society about $43 billion between 2012 and 2016.

In addition, there are about 2,000 incidents each year where trains collide with cars, trucks, and buses in places where rails and roads intersect. These crashes are particularly deadly: for every 100 there are about 40 injuries and 12 deaths, a rate that has stayed dreadfully consistent for the past three decades. Railroads are also inherently dangerous places to work, with lots of ladders, complex movements, and heavy machinery. An average of 14 railroad personnel died on the job annually between 2009 and 2019.

Fortunately, the bipartisan infrastructure package winding its way through Washington has the potential to address parts of the problem, with $5 billion for rail improvement and safety grants, and another $3 billion for grade crossing safety improvements.

But we need to go further.

Currently, there is no multi-agency federal program devoted to rail trespass prevention. Congress should expand funding for organizations like Operation Lifesaver which, among other things, works to educate the public about the dangers and illegality of trespassing on railroad property. Addressing trespassing will require targeted approaches that involve organizations outside the railroading industry, like social services that address homelessness, addiction, and mental health.

The Amtrak crash in Montana was horribly tragic. The NTSB will carefully and soberly examine the site, make a determination of the cause, and present its findings in a way as to prevent future crashes. But we need to keep in mind that the most acute problems in railroad safety are with trespassing and grade crossing incidents, not high-profile accidents. Future federal policy needs to recognize this fact.

The views expressed above are those of the authors and do not necessarily reflect the views of the Eno Center for Transportation.