I disagree, as I posted elsewhere in this thread these annual death rates are drawn from different distributions. The annual suicide rate is normally distributed, in other words the expected number of deaths will be around the same number each year. Terrorism and pandemics are long-tailed meaning a non-negligible probability that the rate shoots up to something much higher than your quoted two per year.

Not going to claim that that money is all well-spent, but terrorism (and security in general) is a rather different problem. A single failure can cost thousands of lives, and huge knock-on effects, as we've sadly seen.