The problem isn't the actuaries, it's the assumptions. For example, NOBODY KNOWS how much health care will cost in 20 years, but that's central to figuring out

The problem isn't the actuaries, it's the assumptions. For example, NOBODY KNOWS how much health care will cost in 20 years, but that's central to figuring out defined benefit obligations (as just one example). Or, what's a good investment rate of return assumption to use? Small differences (e.g. fractions of a percent) compounded over decades lead to huge differences. How do you model black swan events (e.g. 2008 crash)?