Login

0 votes

Say census data for a company is available at the beginning and end of the year then in the month of April, the company acquires a smaller company and starting from the first of April, includes the policies from this small company on their books. I understand that the assumption of linearity of Px(t) is violated but why does this not have an impact on the mortality rates? Surely there will be more deaths after April? Likewise what is the effect (if any) of a large number of policyholders lapsing on their policies at the same time?

+1 vote

Hi Mutinta

Thank you for your question. Of course it has an impact on mortality rates - as you correctly, state, there is an impact on \(P_x(t)\), which in turn impacts the exposure \(E_x^C\), which in turn impacts your mortality rate (since you use the exposure in calculating your rate).

You are also correct in stating that there will be more deaths after April.

But I suspect you also wish to know if there is an impact of this large influx of policyholders on the AGE to which the mortality rate (either your \(\mu\) or q) applies to. Recall that it is the RATE INTERVAL (RI) which allows us to attach ages to mortality rates, so we need to ask ourselves how this large influx affects the RI. Note also that the RI "does not really exist" in calendar time.

Suppose we consider a life-year rate interval. This large influx of new policyholders has absolutely no effect on the exact age at the start of the RI and the average age at death (which is the age in the middle), since the life-year RI cannot be "transposed" into calendar time. Note also that the **assumptions used** to derive the age at the beginning and age in the middle of the RI are not violated by the large influx of policyholders.

Suppose we consider a policy-year rate interval. Again, this large influx of new policyholders has absolutely no effect on the age at the start and the age in the middle of the RI. Again note that we cannot transpose this RI into calendar time. Also, we do not violate the assumption of deaths occurring half-way through the RI, since we cannot say anything on the relationship between the policy anniversary and 1 April. HOWEVER, if I told you that there was a large withdrawal (or alternatively an influx of) policies three-months after the policy anniversary, then there WOULD be an impact on the average age of death. Can you figure out why?

In addition, IF the RI was a calendar year RI, then the influx would have an effect on the average age at death. But we don't consider this case in our course.

The answer to your last question on lapsing at a particular time is precisely the reverse of what would happen with a large influx. No impact on the ages to which your estimates apply (there is no impact on the RIs and the assumptions we use to get the ages in the rate intervals), but there is an impact on the behaviour of the population.

- All categories
- BUS 1003H - Introduction to Financial Risk (41)
- BUS 2016H - Financial Mathematics (49)
- BUS 3018F - Models (69)
- BUS 3024S - Contingencies (54)
- BUS 4028F - Financial Economics (20)
- BUS 4027W - Actuarial Risk Management (33)
- BUS 4029H - Research Project (5)
- Mphil (1)
- Calculus and Pure Mathematics (3)
- Statistics (16)

...