Hi all, I just wanted to check my understanding of lapse risk when considering term assurance contracts. I am getting a little confused as to which direction actual lapses should be compared to the expected amount when considering risk. General risks: The only risk for TA seems to be higher lapses than expected occur early on. i.e. when the asset share is negative (premiums have not made up for initial expenses and mortality costs yet). Aside from this, selective withdrawals can occur later on when healthier lives lapse, however I don't see how this is a risk so long as the asset share is positive? Is this risk saying that if there are lots of selective withdrawals then the policies that remain are unlikely to attain a positive asset share, since the 'good' policies are not there to offset the 'bad' ones? But surely if a block of polices is currently in a positive asset share then it doesn't matter which ones lapse or don't as the profit has been made. Or am I thinking about this incorrectly, is it actually the future positive asset share that will not be offsetting the future negative asset shares from the 'bad' policies? Balance sheet impacts over a specified period of time: Would there need to be more or less lapses than expected to make a position better or worse? If there were expected to be 5 lapses but only 3 occurred, would the impact mainly depend on if the reserve for the ones that lapsed be positive or negative at the time? Would TA reserves ever expect to be negative (if this is allowed by the regulator)? I understand that lapsing with positive reserves is good because it reduces the BEL. Thanks
Let me attempt to answer the first part. On term assurance later duration withdrawal, I think the issue to remember is that the profit margins are very small as it is a competitive market. So, changes in actual experience vs that allowed for in the pricing will be important. The pricing will allow for some selective withdrawal at later durations. If the actual selective withdrawals are much higher than that assumed, then there will be less profits made than assumed in the pricing (all things remaining equal) as less policies will have paid premiums to the end without claiming death benefits. An extreme case to think about: all lives that will live to the end of the policy term withdraw halfway through the policy, leaving only lives that will claim death benefits before their policy expires. The firm will have priced for a better experience than this. The more good lives withdraw in the second half of the policy beyond what the firm assumed, the closer our firm gets to this extreme case. In terms of asset share, an example could help as well. Let's say total asset share across all term assurances is x at a certain duration towards the end of the policy. If all the lives that will remain alive to the end of the policy term withdraw at this point, the x asset share will have to cover all the death benefits for the lives remaining, which will all claim death benefits (and their claims will be much larger than the premiums they will pay till death). There will be no more contributions to asset share from the healthy lives. I think this brings to the fore another important aspect of term assurance, the importance of volume. The business relies on large volumes of policies and few deaths. I hope that's of some help. The second bit I don't think is crucial before the exam, so I have not spent time on it.
Probably also worth thinking about whether you're dealing with a risk impact question or SCR impact question. So far, the discussion has been in the context of risk but if we were talking about the lapse stress in standard formula SCR, then it's the early duration one that matters. Higher lapse at other durations when BEL has become positive would reduce BEL (as less would have to be held as there are fewer policies on the books) with no impact on assets (as there is no surrender value paid out on surrender). So, assets less BEL would increase.
