I half-heard a discussion on the Today programme this week when child
poverty was being discussed. John Humphrey derided the current measure
which is based on 60% of the median income i.e. 60% of the 50th percentile.
(I'll call this the 60/50 point) He said that in a time of recession, the
median would drop, and the number of households below the 60:50 point could
drop. So paradoxically, although we are all getting worse off, the 60/50
measure would indicate fewer households in poverty. (So presumably would
any p/q measure, for all p and q.)
I think he has a point here - but what exactly is the point? And has
anybody analysed this in a systematic way? (Clearly for any income
distribution, the poverty-value is unchanged by any linear transformation,
for all p and q - but what about non-linear transforms? And how to model a
changing population??
Can anybody throw any light on this please?
JOHN BIBBY
PS: Alan Milburn was involved too - see
http://www.guardian.co.uk/tv-and-radio/2011/dec/18/today-child-poverty-alan-milburn
==== *References*
*
*
http://books.google.co.uk/books?id=i1h6g4HXEBMC&pg=PA67&dq=%22measures+of+poverty%22&hl=en&sa=X&ei=kXfrTriMLc6o8APnsoWWCg&ved=0CDcQ6AEwAQ#v=onepage&q=%22measures%20of%20poverty%22&f=false
http://www.ophi.org.uk/wp-content/uploads/Atkinson-1987.pdf
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|