As we've been saying, migration explains the life expectancy puzzle

For a number of years now we've been trying to beat into the general consciousness something very important about life expectancy figures. No one, ever, measures the age at which people born in a particular place die. What is measured is the age at which people living in a place at the time of their death die. This is an important distinction.

Retirement towns (say, Bournemouth, just to use a cliche) have people who have survived to 65 to retire moving to them. This then raises the average age of death there by the leaving out of the figures of those who die before 65. Actual retirement communities - such as exist in parts of Florida say - will have vastly higher average ages at death than intensive care units for premature babies. Put that way it just seems obvious yet all too few manage to grasp the implication.

That implication being that migration matters and matters hugely. Anne Case and Angus Deaton famously found that Appalachian life expectancies were falling. That's also an area going through depopulation, near all of those who manage to go away to college don't return. Yes, those who go to college tend to have longer lifespans than those who don't. It could, of course, still be true that lifespans of all of that original population are falling but we must at least work out whether it is simply this migration of the likelier to be longer lived portion causing the observed effect.

At which point, something from ONS about the UK's regional disparities: 

The fact that, nationally, life expectancy is increasing makes the decrease in the most deprived parts of England and Wales all the crueller. This contrast had already caught my attention when I stumbled across the sentence mentioned at the start of this column. It’s to be found in the section of the ONS report headlined “London has experienced the most rapid increase in newborn life expectancy”. Discussing reasons, the report’s authors remark: “It is also possible that there is a selective migration of healthy individuals from deprived areas to London for employment or other economic reasons.”

What we do about some perceived or measured problem depends upon what is causing it. If it is simply this internal migration then what should, could, or what would we even want to do about it? 

Or as we might put it, we know that a modicum of exercise increases lifespan. What should we do if getting on your bike to look for work is what is producing these longer spans?