|
Community-level relationships between prime age mortality and rural welfare:
Panel survey evidence from Zambia
T. S. Jayne, Antony Chapoto, Elizabeth Byron, Mukelabai Ndiyoi, Petan Hamazakaza, Suneetha Kadiyala, and Stuart Gillespie
International Food Policy Research Institute (IFPRI) and RENEWAL
November 2005
SARPN acknowledges the International Food Policy Research Institute (IFPRI) as the source of this document: www.ifpri.org
|
|
|
[Download complete version  - 276Kb ~ 2 min (17 pages)]
[ Share with a friend 
]
|
Abstract
Governments and development agencies require accurate information on the impacts of increased mortality rates caused by AIDS on the agricultural sector and rural livelihoods. Several previous studies have estimated the effects of prime-age mortality on afflicted households in relation to non-afflicted households. Given that HIV prevalence rates exceed 15-20 percent in many parts of southern Africa, we question whether non-afflicted households are a valid control group in hard-hit communities because non-afflicted households may nevertheless be adversely affected by the mortality occurring in neighboring households.
Using nationally representative household panel data from rural Zambia, we measure the effects of prime-age adult mortality rates on changes in a set of community level welfare indicators. We find that a rise in community mortality rates from zero to 24.4 percent (which is the difference in mortality rates between the 25th and 75th percentile of all 393 communities) is associated with a 5 percent decline in the land area cultivated at the community level. We find little evidence that cropped area is shifting toward labor-saving crops such as cassava in hard-hit areas as is sometimes contended. Other factors related to
agricultural policy need to be considered when examining the impact of HIV/AIDS on the agricultural sector. Most notably, many countries in eastern and southern Africa had formerly implemented state-led maize promotion policies featuring pan-territorial producer prices, major investments in marketing board buying stations, and subsidies on fertilizer distributed on credit to small farmers along with hybrid maize seed. These maize marketing policies in Zambia were either eliminated or scaled back significantly starting in the early 1990s as part of economy-wide structural adjustment programs. These policy changes clearly reduced the financial profitability of growing maize in the more remote areas where maize production was formerly buoyed by pan-territorial pricing, and has shifted cropping incentives toward other food crops such as cassava (Jayne et al).
We also find relatively small independent effects of prime-age mortality on community indicators of crop production, income, and income per capita. The effects of mortality appear to be complex in that they depend importantly on initial community conditions such as the level of mean education, wealth, connectedness with markets and infrastructure, and dependency ratios. In the case of changes in community crop production per hectare, prior mortality rates have a greater impact than more recent mortality rates, potentially indicating strong lagged effects. In general, the findings of this study offer limited support for the view that prime-age mortality is decimating agrarian-based economic systems in regions that are hard-hit by the HIV/AIDS pandemic. However, this in no way implies that great hardship is not being wrought by the disease in ways that are not measured in this data, and further research from other areas and/or time periods will be necessary before strong conclusions can be generated about the effects of AIDS-related mortality in regions of the world where prevalence rates are very high.
Introduction
A growing literature has focused on understanding the effects of the HIV/AIDS pandemic on rural livelihoods and the agricultural sectors in Africa (Ainsworth, Fransen, and Over; Barnett and Whiteside; Gillespie and Kadiyala; Mather et al.). In some parts of southern Africa, HIV prevalence rates are as high as 30 percent among individuals between 15-45 years of age. Several nationwide household panel surveys from relatively hard-hit countries (Kenya, Tanzania, and Zambia) indicate that, over a 3-year survey interval, roughly 6-10% of rural households suffer one or more disease-related deaths of a primeaged individual (Yamano and Jayne; Beegle; Chapoto and Jayne), and there is overwhelming evidence that most of the mortality in these age ranges are related to AIDS (Ngom and Clark).
However, efforts to accurately estimate the economic impacts of AIDS-related mortality are fraught with difficulties. To date, almost all of the quantitative micro-level studies have studied the effects of mortality at the household level, even though it is likely that mortality shocks are transmitted across households. This situation, in which a relatively small percentage of households incur a shock, but the shock is spread across households in a community presents methodological challenges for estimating the full effects of the shock using household survey data. Most household-level panel studies, using difference-in-difference, household fixed-effects, or random-effects models, have measured the effects of mortality in afflicted households on differenced household-level outcomes, typically over a 2–5 year time frame, compared to differenced outcomes on nonafflicted households. Yet if non-afflicted households are likely to be indirectly affected by the mortality occurring around them, non-afflicted households may not be a valid control group. In communities hard-hit by HIV/AIDS, households not directly incurring a death may nevertheless be affected by taking in orphans, losing access to resources owned by kin-related “afflicted” households, intra-household resource transfers to afflicted households, and broader effects of high mortality rates on communities’ economic and social structures. To date, little quantitative economic analysis has attempted to measure the effects of mortality on rural welfare other than at the household level.
This study measures the effects of prime-age mortality on rural welfare using the “community” as the unit of observation. Data is drawn from a panel of 5,420 households surveyed in 393 communities in Zambia in 2001 and 2004. We compute community-level
adult mortality rates from household data along with mean household welfare indicators for all communities. OLS difference models are estimated to measure the relationship between mortality rates and indicators such land cultivation, crop output, and per capita income at the community level, controlling for time-invariant unobservables and initial community conditions. This study is the first approach, as far as we know, to estimate the impacts of AIDS-related mortality on entire rural communities, including afflicted and non-afflicted households alike, using micro survey data. The findings should be important for governments and development agencies especially in southern Africa where HIV prevalence rates are the highest in the world and where the full impacts of the disease remain largely speculative.
|
|
