Monday 29 September 2014

Zimbabwe is food secure this season, but more questions raised

The annual ZimVac assessment based on a national sample survey of over 10,000 households and carried out in May came out a month or two back. Unlike last year, when alarm bells were rung over a potential food security catastrophe, this year the prognosis was good. Excellent rains, including in some of the drier and usually more food insecure parts of the country, resulted in a bumper harvest.

Last year I critiqued the use of the headline figure from the assessment as potentially misleading. The same limitations of the survey apply, but the media reporting is more balanced this year (with some extreme exceptions – see comment string in an earlier blog). The survey is based on the 2012 Zimstat sampling frame and covers a large number of enumeration areas across the country, sampled proportional to population densities. Annoyingly the report still doesn't separate out communal areas and resettlement areas, and my guess is that there remains some sampling bias. More on this below. Last year fortunately the dire predictions were not borne out. In part this was because the rains came, and a green crop filled the hunger period, but also I hypothesised in an earlier blog that the production from new resettlement areas was being undercounted. I suspect this remains the case.

Anyway, I thought blog readers would like a quick summary of the report, as without an impending disaster the media has largely ignored it. You can read the powerpoint report in full, which covers all sorts ranging from nutrition to sanitation. I will concentrate on agricultural production and food security, and draw text directly from the report.

The Ministry of Agriculture, Mechanisation and Irrigation Development estimates that the country will have a cereal harvest surplus of 253,174 MT in the 2014/15 consumption year from a total cereal harvest of 1,680,293MT.

Maize remained the major crop grown by most households (88%) compared to 80% for 2012/13, while groundnuts were the second most grown crop. Generally, the proportion of households growing crops increased except for cotton which showed a decline (due to the collapse in prices) and soya beans which remained unchanged.

Nationally, average household cereal (maize and small grains) production was 529.5kg. This was higher than last season (346kg). In Masvingo maize production averaged 339.7 kg and small grains 126 kg, given a total of 525.7 kg per household. Overall, average household cereal production was highest in Mashonaland West and lowest in Manicaland, and the contribution of small grains to total household cereal production was significant in Masvingo, Matabeleland North and Matabeleland South.

While improvements, these average figures are still low. And compared to the production levels from new resettlement households minute. Our studies in Masvingo, even in the poor rainfall years of between 2010 and 2012, show much higher averages (although with variations). Gareth James' studies from Mashonaland shower higher outputs still. Again in the poorer rainfall years, he recorded average outputs of maize some 12 times these average national figures for all cereals for the good rainfall year of the past season. Of course the new resettlements have proportionately fewer people and so appropriately in a national representative sample this should be reflected. But, without data broken down and without indications of variation, the ZimVac study still fails to capture this story. As I have argued before (many times!), this is important for policy, and for thinking about national food security.

The ZimVac survey showed that for the 2013/2014 agricultural season approximately 45.2% of the households benefited from the Government Input Support Scheme, which was the main source of inputs. The proportion of households accessing maize inputs through purchase remained unchanged (39%) from 2013. About 2.3% of the households accessed their maize inputs from NGOs which was a decrease from 4.0% in the 2012/13 season.

Given the higher levels of production, the national average maize price was $0.37/kg down from $0.53/ kg during the same period last year. This pattern was also reflecting at the provincial level. Matabeleland South recorded the highest maize price ($0.65/kg). This was the same pattern during the same period last year.

Livestock (cattle, sheep and goats) were in a fair to good condition when the survey took place. Grazing and water for livestock were generally adequate in most parts of the country save for the communal areas, where it was, as is normal, generally inadequate. However, the report notes, there are marginal parts of Matabeleland North and South, Midlands, Manicaland and Masvingo provinces which had inadequate grazing which may not last into the next season.

According to the report, around 60% of the households reported not owning any cattle. Mashonaland East had the highest proportion of households not owning any cattle and Matabeleland South had the least. Nationally, only 14% of the households owned more than 5 cattle with Matabeleland South and Matabeleland Matabeleland North having a higher proportion of households owning more than 5 cattle.

Like the cereal production data, these national and provincial figures are very different to what we have found (and Gareth and others) in the new resettlements. Here cattle ownership is far higher, reflecting the richer, more capitalised form of farming found. Of course the ZimVac study may suffer from under-reporting, as in many large-scale surveys with huge samples, but the contrasts are interesting – and again potentially important.

In terms of food consumption, Masvingo had the highest proportion of households consuming an acceptable diet (75%) and Matabeleland North had the lowest (54%). This showed increased local availability of foodstuffs, and improved off-farm opportunities. However, nutritional indicators remained low, including a high prevalence of stunting. As commented on before, this mismatch between food intake and nutritional indicators remains puzzling.

So, following the food balance methodology the assessment adopts (see discussion of the methodology and its limitations in an earlier blog), the report estimates that for the 2014/15 consumption year at peak (January to March next year) is projected to have 6% of rural households food insecure. This is a 76% decrease compared to the (disputed) estimate the previous consumption year.

This proportion represents about 564,599 people at peak (which may of course be people suffering deficits for only a few days), not being able to meet their annual food requirements. Their total energy deficit is estimated at an equivalent of 20,890MT of maize; actually a very small amount, and not suggesting any urgent need for food aid, given the margins of error in the estimates. Matabeleland North (9.0%), Matabeleland South (8.3%) and Mashonaland West (7.7%) were projected to have the highest proportions of food insecure households. By contrast, Manicaland (2.7%) and unusually Masvingo (3.4%) provinces were projected to have the least proportions of food insecure households.

So in sum, a good harvest results in a good food security situation. This is of course good news, and no surprise. But the report and the analysis still raise many questions. I hope that those working on food and farming in Zimbabwe can join forces and think harder about questions of sampling, the contributions of the new land reform areas to production, and the complex dynamics at the heart of the food economy that underpins food insecurity prevalence and distribution. The ZimVac annual survey is a major contribution, but with some thought and adaptation it could be contributing much more to our understanding of changing livelihoods and food economies in the post-land reform era.

This post was written by Ian Scoones and originally appeared on Zimbabweland