Graham’s preferred course requires a tremendous amount of blood and treasure. Among the most obvious questions it raises is, “Would it be better to refrain from risking that blood and to spend that treasure elsewhere?” And Graham’s interventionist streak is not limited to Afghanistan. He favors fighting with Saudi Arabia in Yemen and an indefinite troop presence in Syria and Iraq––and war in Iran and North Korea if they do not accede to U.S. demands about their nuclear programs.

In each case, he argues that American deaths could result if his advice is not taken, but fails to contend with opportunity costs. He raises the possibility that another 9/11 could occur if the U.S. leaves Afghanistan, but never the possibility that, for instance, staying in Afghanistan instead of redirecting that money to efforts to secure fissile material most vulnerable to theft could end with a dirty bomb in New York City.

By spending time with Trump, flattering him at times, and appealing to his aversion to anything associated with Obama, Graham may be succeeding in making him more interventionist than he would otherwise be. But until Graham’s public arguments are stripped of the sunk cost fallacy and incorporate the opportunity costs of intervention, they should be rejected by observers more clearheaded than the president as unlikely to improve foreign policy or to make Americans any safer.