Zimbabwe experiences severe drought every two to three years. In intervening years, parts of the country are periodically affected by floods. Correspondingly, the country frequently hosts drought or flood relief programs targeting the recovery of smallholder agriculture. The most common programs, involving the distribution of seed and fertilizer, have been implemented in one or another part of the country during at least ten of the 24 years since the country achieved its independence in 1980.
Zimbabwe experienced severe drought once again during both the 2001/02 and 2002/03 cropping seasons. Further, the impact of these recent droughts was measurably worsened by a rise in unemployment, high (100-500 percent) rates of inflation, a decline in gross domestic product, and an estimated 26 percent rate of HIV/AIDS incidence among adults. Further, maize import and price controls contributed to severe shortages of grain on both urban and rural markets. In past years, farm households have responded to drought by increasing their food purchases. In 2003, it was periodically difficult to find grain for purchase. Consequently,
households were assumed to be more likely than usual to consume their limited seed stocks.
Despite the frequency of agricultural relief programs, little is know about their efficacy. Seed distribution is assumed to contribute to an expansion of cropped area. But it is difficult to find independent data measuring such gains. Fertilizer is assumed to increase production levels and productivity. But most relief programs simply assume these gains. Nonetheless, each year drought re-occurs, these programs are simply started afresh.
This study re-examines these assumptions. The analysis summarizes the results of three major farm surveys designed to assess the distribution of seed and fertilizer inputs following the 2002/03 season drought in Zimbabwe – during the 2003/04 cropping season. The analysis reveals that while the relief seed and fertilizer were generally well used, there remain substantial opportunities for improving the effectiveness and impact of these input distribution programs.
The targeting of households destined to receive relief needs improvement. While many of the non-governmental organizations (NGOs) distributing inputs identified explicit criteria for the selection of needy households, these lists were difficult to implement in practice. In consequence, there was little difference in the poverty levels of households that received relief inputs compared with those that did not receive these inputs. Many NGOs tried to target households affected by HIV/AIDS. Yet households with orphans, or female headed families were just as likely to have received relief inputs as male headed households or those without orphans.
Problems were also widely apparent in spatial targeting. Almost 15 percent of households received input packages from more than one non-governmental organization (NGO). In some districts, more than 25% of households received similar packages or relief inputs from multiple NGOs.
Targeting can be improved through better information sharing about the regional distribution of production losses and the spatial distribution of NGO activities. In addition, simpler proxy variables are needed to identify poorer households. One such
proxy that appears robust in much of Zimbabwe is the ownership of cattle or donkeys for animal traction.
The distribution of seed does not appear to have contributed significantly to the expansion of cropped area. Instead, much of the relief seed appears to have replaced stocks available on local markets. This includes seed saved by many households from
their previous harvest. Despite shortages of grain on the local market, and despite two consecutive years of drought, many households were still able to retain seed stocks.
Farmers appear to have benefited most from the distribution of new, improved varieties. This was the first season in more than two decades that relief agencies were allowed to distribute open pollinated maize varieties. While virtually all smallholders
had adopted hybrid maize, the rising costs of this seed in recent years had led many to replant seed derived from their previous season’s grain production. This was contributing to a decline in average maize yields. The delivery of open pollinated
varieties offered farmers a cheaper, more sustainable, alternative. But unfortunately, most farmers receiving this seed had no idea whether they were getting hybrid or open pollinated varieties. Major investments are now required to teach farmers about the
differences between alternative varieties.
The survey results also indicate that greater care needs to be taken to assure relief seed is of good quality. Much of the seed, especially for crops other than maize was of questionable origin. A significant share appeared simply to be grain cleaned to seed specification for physical purity and germination. Farmers asked why they were receiving varieties they already owned. In at least two cases, seed of poorly adapted varieties was imported and distributed to farmers. This produced limited yields late in
the production season (which fortunately was prolonged by late rains). In these cases, the recipients of relief seed would have been better off planting seed available on local markets. These problems were worsened by poor and incomplete seed labeling, and
in some cases, wrong labeling.
The diagnosis of these problems has led to the drafting of a relief seed protocol calling for better labeling and the promotion of known varieties. These interventions alone, however, will not resolve the problem of shortages of important seed crops of limited interest to commercial seed companies. One additional solution is to encourage the establishment of seed security stocks for promising new varieties.
The study reveals that substantial gains in production and productivity were derived from the targeted application of small quantities of chemical fertilizer. More than 150,000 farmers received 25 kg of ammonium nitrate (AN), most with information
about how to apply this in the form of micro-doses to a grain crop. Associated demonstration trials on more than 1400 fields highlighted an average yield gain of 30- 50% derived from the application of only 10-20 kg N per hectare (about one-quarter
the commonly recommended levels). Importantly, these gains appear consistent across regions and farmers. In effect, small doses of nitrogen-based fertilizer appear to offer much higher returns than the delivery of seed – particularly if this seed is of
Unexpectedly, the major determinant of the area planted by poorer households was not the availability of relief seed, but access to draught power. Families owning cattle or donkeys planted 60 percent more land than those without. This is linked with an 80 percent average increase in grain harvests. These results suggest the need for introducing labor saving tillage systems, or special programs to resolve draught power constraints.
Larger gains can also be achieved by strengthening the technical assistance provided with these relief programs. Less than one-quarter of the recipients of relief inputs received any kind of extension advice. And the majority of these extension contacts
occurred only once. Most farmers, correspondingly, could not identify what seed varieties they received – even if variety names were printed on the bags. An opportunity to educate farmers about new seed and production technologies was lost.
Finally, the study initiated an examination of the relative benefits of alternative distribution strategies, including a) direct input handouts to farmers, b) the distribution of seed and fertilizer using vouchers redeemable for designated input packages at rural retail outlets, and c) the distribution of vouchers redeemable at seed fairs. Direct handouts appear the easiest delivery method, but also the strategy most disruptive of rural markets. Rural retailers have little incentive to stock seed or fertilizer if they suspect an NGO will be handing out these inputs for free. The vouchers redeemable at local shops offered retailers a marginal payoff, but did not improve incentives to stock agricultural inputs because the redemption packages were pre-defined. In both cases, the predetermination of input packages was linked with the distribution of seed that was never planted. The seed fairs offered more choice, and generated income for local communities. A larger share of the seed derived from the fairs was planted. But the high seed prices offered in order to attract seed may have undermined the operation of traditional village markets. And questions remain about the quality of some of the seed flowing through these fairs. In general, more experimentation is warranted with voucher type schemes linked with the development of rural input markets.
Overall, the evidence strongly suggests that agricultural relief programs need to move away from an emphasis on handouts to encompass the pursuit of more explicit development goals. These programs may still target subsidized assistance to poorer households most severely affected by poor rains or socioeconomic constraints. Yet many of these households are likely to remain chronically poor unless they are more methodically assisted with improved varieties, better extension advice or strengthened
markets. They will continue to need assistance during the next drought, and that following two years later. Larger, more sustained gains can be achieved by improving the quality of assistance, rather than concentrating, as most programs do today, on the
numbers of households assisted, and the numbers of input packages delivered.