Mortality and morbidity (2) (Selection (Five types of selection (STATiC),…
Mortality and morbidity (2)
This is the process by which lives are divided into separate groups so that the mortality/ morbidity within each group is homogeneous. In other words, so that the mortality/ morbidity of all the lives in the group can be modelled using the same stochastic model.
In practice, however, the word 'selection' can also be used to refer to the source of heterogeneity itself, i.e. the characteristic responsible for the differences in mortality/ morbidity.
The word 'selection' can also be used in the context of anti-selection. That is, an individual taking advantage of inefficiencies in a provider's pricing basis to secure better terms than might otherwise be justified.
Five types of selection
When homogeneous groups are formed we usually assume that the factors used to define each group are the causes of the differences in mortality observed between the groups. However, there may be other differences in composition between the groups, and it is these differences rather than the differences in the factors used to form the groups that are the true causes of the observed mortality differences. Ascribing mortality differences to groups formed by factors that are not the true causes of these differences is termed spurious selection
mortality improvements that are actually due to increasing the strictness of underwriting
Geographical mortality differences that are actually due to a different balance of high and low risk occupations
Within a population, mortality and morbidity rates normally vary over calendar time, essentially due to medical advances. This effect is usually observed at all ages and the usual pattern is for mortality rates to become lighter over time (although there can be exceptions). This is called time selection.
individuals living 30 years ago experienced heavier mortality than lives of the same age today
individuals with life insurance policies with a sum assured exceeding £150,000 are more likely to have taken out the policy recently
This is any selection that leads to an adverse effect on another party (for example, on an insurance company). It usually involves an element of self-selection, which acts against a controlled selection process which is being imposed on the lives.
people who decide to purchase an immediate annuity with pension funds usually experience lighter mortality than those who decide not to do so.
People who smoke will tend to seek life insurance from companies who charge identical premiums for smokers and non-smokers, whereas non-smokers will apply to companies who differentiate between them and therefore charge cheaper premiums to non-smokers. The first company will suffer from adverse selection, as the ratio of smoker to non-smoker lives that it takes on will increase.
Temporary initial selection
It is usually the case that mortality/ morbidity rates depend on the duration since some event, up to some duration
. After duration
they are independent of duration. This phenomenon is called temporary initial selection, and
is called the length of the selection period,
Temporary initial selection occurs in life insurance polices that have been underwritten, as the mortality rates of lives who have recently taken our policies are lower than those of the same age who took out policies several years ago.
Class selection occurs where each group is categorised by the nature of a particular characteristic of the population
gender (categories male and female)
occupation (categories manual and non-manual employment)
Variations in mortality between different subgroups are most noticeable at working ages
For older populations, mortality in retirement is less variable between different groups (although the evidence is disputed and detailed analysis is complicated by low volumes of data at the highest ages).
This convergence of mortality between subgroups at higher ages is referred to as mortality convergence.
How decrements can have a selective effect
One way in which lives in a population can be grouped is by the operation of a decrement other than death. Those who do and those who do not experience this selective decrement will experience different levels of mortality or morbidity. Examples of selective decrements include:
Those getting married usually experience lighter mortality and morbidity than those of the same age who do not get married.
Ill-health retirement results in lighter mortality among the remaining active members of a pension scheme. This is sometimes called the 'healthy worker' effect.
Withdrawal from a pension scheme can be associated with voluntary resignation. This tends to select those with lighter mortality.
Term assurance policyholders who lapse are likely to have lighter mortality than those who keep their policies in force.