From: Matt
Sent: 10 May 2001 01:00
Subject: Re: Life tables
Dear Carl

Thanks very much for these and your time yesterday  -  it was good to talk to you.

I have wondered whether commercial (pensions and insurance) actuaries may have information that is relevant as regards differential mortality and income in industrialised countries.   More generally, it seems that to tackle these questions I need a synthesis of at least:

-   welfare (macro)economics
-   "micro"  -  level observations
-   theories of utility
-   population studies.

What I am attempting overall is to examine rigorously the process by which conclusions are drawn as to increases or decreases in welfare, from data collected on people living at different times, in any population where survival rates are highly variable.    This includes looking at non-linear relationships between income/consumption and welfare (taking into account changes in mortality risk), changes in ratios of earners to non-earners, choice of weightings, choice of average, and estimates of probability for inferences.   Some of the questions I have relate to the way that progress is measured on the international development goals, particularly poverty reduction, and to possible causal relationships between rates of progress on the different goals.

My own background, by the way, is in classics and experimental psychology, and my instincts on micro-level trends in poor countries come partly from living with ordinary people in Bangladesh in the 1980s.   I believe that it is possible for social science to come to more reliable conclusions about aggregate trends in welfare in poor countries, and to narrow the gap between statistical methods and observational methods;   I also believe that this cannot be done without rethinking some aspects of methodology from first principles.