Multiple decrement versus associated single decrement probabilities
When developing the multiple decrement life table, we have argued that cause-specific probabilities of dying (nqxi) must add up to the all cause probability of dying in an age interval. Imagine a world with two causes (i and –i). Everyone can only die from one of those causes and that cause is in principle observable:
The relationship between associated single decrement probabilities and all-cause probabilities is a bit more complex. This is illustrated by the venn diagrams below. In the absence of other causes, cause I would be responsible for the deaths represented by the venn diagram on the left, i.e. it would claim casualties that in reality have succumbed to other causes (the shaded area):
Therefore: