Still spit-balling – I definitely could be wrong…
But I’m thinking, based on the big bump in mortality rates often observed in the first 3-5 years, that the anti-selectors are more so critical/terminal than chronic illness. The latter I’d agree could be materially affected by mortality improvement. My thinking is that the causes of mortality improvement (short of some major cure for cancer, etc) is generally something that more so emerges gradually over time and primarily affects people before or at the early stages of illness or even “pre-illness” compared to somebody who is at the later stages of disease. Of course there would be some improvement (eg drugs that extend life, etc). But I’m guessing that improvement comes more from avoiding illnesses and early detection more so than curing/treating them at critical stages.
My “overwhelming” thought is due to the size of the big bump in early years. Even if non-anti-selectors have improvement, it is overwhelmed when measuring early year claims as a % of total mortality rates. The good news (if your company sells enough to be credible) is that you may be able to detect improvement in the early years as you study early year mortality. If not credible, maybe your reinsurers have opinions on that.