I don’t think so.

Consider 200 identical companies which survive or fail together each year (yes, it’s artificial of course). It’s quite possible for the probability that a given company fails in a year to be 1/200. In this case “in the long run” you’d expect 199 years to have no failures and 1 year to have 200 failures. In a layman’s sense you don’t “expect” one failure per year.

Probabilities as low as this are notoriously different to judge, especially when the past is such a poor gudie to the future. In theory regulators (rather than firms) are looking at this sort of dependence issue via “macro-prudential regulation” programmes. We’ll see!