From Matt Berkley 
Subj:	Epidemiology and statistics
4 October 2002
To:  cmurray@hsph.harvard.edu

Dear Professor Murray

I think you may be interested in the following.

I am currently writing about poverty statistics   -   in particular the relationships between life length and population averages.

If the poorest live longer, then average income may fall, and also the proportion of poor people may rise, as a  result.

Perhaps this kind of effect has been considered in relation to epidemiological data   -  if there are more deaths from AIDS or malnutrition, then they may contribute to a lower prevalence level.   

The state of knowledge on this kind of thing in economics is basically nothing, and there is no theory dealing with even starting to think about it.   


[Note, 16 April 2018:  I was perhaps mistaken.  

I was not aware of, and am still not aware of, any reply from Jeffrey Sachs to my explanation of the issue to him on 11 April 2011.   It is possible that I missed a reply.  In my email, I had suggested the issue might be of interest to the WHO Commission on Macroeconomics and Health which he was chairing.  

In December 2001 I was aware of the publication of his main CMH report.  Its main headline was about the economic value of saving lives.  But I may not have been aware of the accompanying paper by Jamison, Sachs and Wang also of December 2001.  This spelt out basically what I had told him in April, and confirmed that it was an important issue in relation to AIDS - economists had made a large mistake.  

Perhaps I did not appreciate that even the main headline of the CMH report was in line with what I had explained to him, and that this - like the AIDS paper which I may have only seen much later - was very different from the previous approach of CMH output.

The CMH paper dealing with AIDS was perhaps more explicitly similar to what I had explained to Jeffrey Sachs than the main report.  

In any case, there was already some relevant - though long-neglected - economic theory on valuing life length.  This had not been applied much or at all to poverty.  

As well as the Sachs team, Ravi Kanbur - to whom I had explained the issue the week after I contacted Sachs in 2001 - had raised it at a conference in spring 2002.  He may have been working on some theory by the time I wrote this email in October.  Bourguignon and Morrison produced a paper in 2002 using country-level life length estimates to infer global inequality changes.  I do not know whether Bourguignon, as incoming Chief Economist of the World Bank, had been advised of the issue by anyone else with whom I had raised it, such as Oxford professors or the head of the World Bank's World Development Indicators project.]


Conclusions as to average benefit to poor people from policies which raise average income or lower the proportion of poor people may be in doubt.    But it is hard to know whether such effects have up to now been large or small.    In the case of economics, the presence of significant effects of differential mortality -  and related effects of demographic change, such as changing ratios of adults to children   -    could provide a partial explanation of different rises in GDP in different countries.    I would be interested to know any
estimates of this kind of effect in the field of epidemiology.   

Many thanks.