We’ve pointed this out a number of times before. No one - absolutely no one at all - measures life expectancy by place of birth:

It is a league table that no one wants to top. For the first time in 20 years, Blackpool, a once-glamorous seaside resort, this week overtook Glasgow to have the lowest average male life expectancy in the UK.

Men born in Blackpool will now live until just after their 73rd birthday on average, according to the Office for National Statistics (ONS) study, six years less than the average in the rest of England.

The figures highlighted an uncomfortable truism about modern life in Britain: wealth brings health and poverty kills, in what the ONS called a “clear” north-south divide.

The first paragraph could well be correct. The second is not. Yea even the ONS is wrong on this:

The 10 highest local area male life expectancies at birth were all located in the south of England; the 10 lowest were in Scotland, in the north of England and in Wales.

We do not measure lifespan by place of birth. We measure life span by place of death.

To give a personal example, as and when - hopefully long delayed - the Grim Reaper comes for me I will contribute to life span figures for a small town in the Alentejo, in Portugal. Not my place of birth (Torquay), my place of upbringing (Bath, relevant for any measures about Sure Start, primary school and all that) or any of the other places I’ve lived (the US a couple of times, Italy, Russia and so on). Now, yes, I’m probably more mobile than most over life so far. But it’s still vital to grasp this point.

At the level of detail the ONS is reporting there are some 400 (-ish) local areas in the UK. What is being measured is age of death of a person then registered against the one of those local areas that death happened in. OK, useful number. Interesting to know. But what it is not is the expected lifespan of someone born in that local area.

For people move around during their lives. Few move country quite as much as I have done but there’s always a certain amount of internal migration. And the lower we go in our area measurement the greater will be that migration rate. It’s almost certainly true that the vast majority of deaths in the UK happen to someone born in the UK. This becomes less so when looking at the constituent nations, less again when looking at county, lower again at this local area and who, really, thinks that anything more than a distinct minority dies in the same council ward they were born into?

The importance of this is that without knowing the rates of internal migration - which we do not know - we cannot now ascribe conditions at birth to length of lifespan. Sure, sure, we all agree there will be influences. Obviously. But our expected lifespan numbers are a proxy for this. And a proxy that does not work at the more detailed levels of population study. Therefore we cannot use the more detailed numbers for smaller geographical areas as guides to policy.

Anyone who now tries to say that we’ve got to do something about babbies in Blackpool is not just wrong they’re ignorant. Because the information we’ve got doesn’t tell us anything about babbies in Blackpool. It tells us the age people in Blackpool die at, but it’s “people in Blackpool” not “people born in Blackpool”. Those two are only the same if there’s no migration at all across local authority boundaries in and or around Blackpool. Something that simply is not true.

Yes, yes, statistics are nice to know and all that. But before we use them to do anything we need to know exactly what is being measured. Expected lifespan statistics are the age at which people die in a particular location. Not, not at all, the age at which someone born there is going to die. We simply do not have, in any shape or form, statistics relating lifespan and place of birth at any detailed geographic level.

Tim Worstall