Griffith Feeney, Population and Development Review 27(4):771-780 (December).
Available on JSTOR. If you want a reprint and don't have access to JSTOR, email me.
In June 2000, an estimated 25 percent of adults in Zimbabwe were living with HIV/AIDS. Statistical data on the impact of the epidemic, though problematic in many ways, are better for Zimbabwe than for many other countries in sub-Saharan Africa. This analysis presents estimates of adult mortality in Zimbabwe based on multiple sources, including registered deaths adjusted for incomplete reporting, estimated at approximately 50 percent. Comparison of estimates from different data sources shows that they are subject to substantial errors. At the same time, the estimates leave no doubt that adult mortality risks in Zimbabwe more than doubled between 1982 and 1997. The evidence that this rise is due to AIDS deaths is circumstantial, but very strong; there is no credible competing explanation.
While doing this work on adult mortality in Zimbabwe I developed a modification of the Brass “generalized growth balance” method that addresses the objections raised by Trussell and Menken. The modified method would have been difficult to implement by the hand calculator or main frame computer methods in use at the time, but it is easily implemented on a spreadsheet. The following files provide information on the modified method.
The Brass and Trussell-Mencken papers in the Asian and Pacific Census Forum are unfortunately not readily available. Fortunately, however, they are also not restricted by copyright and so may be provided here. These scans were made some years ago and are not of very good quality. I will replace them with better quality scans when I can get good copies of the originals and rescan them with newer equipment.