via Food crisis in Zimbabwe: 2.2 million at risk. But where do the figures come from, and what do they mean? | zimbabweland by Ian Scoones 16 September 2013
[Read the original at ZimbabweLand for comments and additional context and links]
The newspapers have been full of commentary on a looming food crisis in Zimbabwe. This has followed from the World Food Programme’s press release that 2.2 million people will be in need of food aid in the coming months. The Commercial Farmers Union has called it a ‘man-made crisis’, the direct result of the ‘chaotic’ land reform, and a decade of inappropriate policies.
I wanted to find out a bit more about where the 2.2 million figure came from. It’s a big number, and would mean a lot of food imports, way beyond the means of the Finance ministry. After a bit of digging I eventually found the figure, buried on page 122 of the ZimVac (Zimbabwe Vulnerability Assessment Committee) livelihood assessment draft report for 2013.
Each year ZimVac, a coalition of NGOs, researchers and government agencies, undertake a major rural livelihood assessment, based on a sample of over 10,000 households across the country. The sample is drawn according to the latest ZIMSTAT ‘master sampling frame’, and the resulting data is aimed to be representative of the country as a whole. It’s an excellent and important initiative, but it has its deficiencies, as those involved readily admit.
The process for deciding the headline figure is complex. It involves assessing for each household all the cereal production, and then adding in income from employment, remittances, livestock sales, and other sources of income that could be used to buy food (p. 120). Assumptions on prices and market availability are used to translate income into food and in turn energy (p.121). The food security assessment is based on the household’s potential access to enough food from all sources, including purchases, to give each member a minimum of 2100 kilocalories per day in the consumption period 1 April 2013 to 31 March 2014 (p. 119). The total number in food deficit figure is then calculated as a sum of all of those experiencing any negative balance in the accounting period.
It’s a complicated procedure with lots of steps and plenty of assumptions. What the headline figure doesn’t indicate – although the report does, and the background documents for the ZimVac surveys over the years are quite transparent about this – is that the big number includes many people who may have a projected deficit for actually a very short period. Indeed, at the time of the survey in May 2013, over 80% of households surveyed had no hunger problems with only a very small proportion recording ‘severe hunger’ (p. 115). The report shows that there is a progression of food insecurity, with a peak of 2.2m people expected in January to March 2014 (p.124). 31% of the total (683,000 people) move into food deficit only in this crunch period before the next harvest; and some of whom may in fact be food insecure for only a very few days.
The 2.2 million figure is of course a good flag-waving number for the WFP to raise funds, and for the CFU to bash the government for the land reform (and even President Mugabe is now joining the critique of the ‘new farmers’), but the actual implications are more complex. Here are five reasons why we need to be cautious about the figures.
- First, there’s geography: as the report shows the problems are concentrated in the dry south of the country which experienced the worst season in terms of rainfall and its distribution (p.125-6).
- Second, there is almost certainly (as ever in surveys) an underreporting of income, and so purchasing power. Since in drought years, market purchases are essential for food entitlements, this is rather crucial.
- Third, the assessment model allows for only limited sales of livestock to compensate for food deficits (households are assumed to retain a minimum of 5 goats and 3 cattle). Yet livestock is precisely the asset in the drier parts of the country that are used in times of drought to exchange for grain, and distress sales are common, and important for food security.
- Fourth, remittances are especially important in drought-prone areas, yet the figures used in the model for this year are based on recall of last year’s receipts. Last year was of course a relatively good year for rural production, and so remittance flows inevitably dropped. But this year, you can be sure, they will increase in response to the shortfalls. For perfectly good reasons, the model does not account for this, but it’s another reason why we can expect things to be not as bad as predicted.
- Finally, the assessment does not include early cropping – for example of green maize – which is often important in that crunch period before the ‘proper’ harvest.
For all these reasons and more, we should be cautious about the headline statistics, and understand in more detail what happens to whom and where.
One of the most striking figures in the report is the prediction that 98% of rural households nationally will hit a food deficit by next March if only cereal production and stocks were included (p. 123). Of course this includes those with no food production to speak of, such as farmworkers and other rurally-based non-farm households. But even discounting this group, this is striking, and does suggest a problem in agricultural production, as Charles Taffs of the CFU indicates. However, again we must be cautious in jumping to conclusions.
One big concern I have with recent national surveys is that they have been sampling according to old sample frames set before the land reform. This was the case for the 2011 PICES (Poverty, Income, Consumption and Expenditure Survey) study and the 2010-11 Demographic and Health Survey, both using the 2002 census sample frame. I have been assured that the ZimVac survey for 2013 used an updated sample, with ‘enumeration areas’ allocated proportional to population distribution derived from the 2012 census. If so, this would have included the significant populations, especially in A1 areas, who are – at least according to our data from Masvingo – producing more and doing better than their counterparts in the communal areas, where most the earlier rural samples are drawn from. And in our study areas on A1 sites we see between half and two-thirds of the households producing sufficient cereals for the year – not just 2%,
Following the 2012 census, ZIMSTAT is revising the national ‘master sampleframe’, and hopefully from now on national surveys will be statistically more representative. Unfortunately it is still difficult to stratify the data according to land use types, and so distinguish between resettlement areas and others, so Taffs and co should probably hold off on their outright dismissal of land reform on the back of this data for now. As ever, it’s more complicated than it first seems.
That said, last season was unquestionably a worse one than experienced in the last few years, including in Masvingo. It also hit some higher potential areas hard, with a very unevenly spread rainfall. Despite improvements since 2009, input supply was again erratic and untimely last year. Also, maize area planted was again down, reflecting the shift from food crops to tobacco in some areas, perhaps especially in those food producing areas in the higher rainfall zones. This restructuring of the crop system is directly driven by incentives – tobacco, supported through contract arrangements – , is a much more profitable crop than maize, especially if marketed through the Grain Marketing Board. Over the last decade or more we have seen switches to small grains (although plantings were down this past year according to ZimVac), but these are still a small percentage of total crop output, and it remains maize that drives the food economy, although much of this circulates outside the formal channels, and so isdifficult to capture in national statistics.
So what should we make of all this? Certainly there is going to be a problem of food deficits in the coming months. However, problems are going to be concentrated in a certain time period, and outside a few areas and for more vulnerable people, it’s not going to be as bad as the headline figure and the media commentary perhaps suggests. Imports will certainly be needed, and targeted food aid will be important, but other coping strategies will also come into play to offset the worst.
Indeed this seems to have been the pattern over many years now. There is a ritualised flurry of activity around this time of year, with the aid agencies calling for funds to support food aid, and those critical of land reform saying that this ‘proves’ that Zimbabwe has gone for food producer to ‘basket case’. Yet by the end of the season, the expected famine has not occurred and, although hardships unquestionably are faced, the scale and depth of the problem is not as expected. This can be explained due to both sampling and non-sampling errors inherent in the standard surveys; but also significantly because assessments have not got to grips with the new patterns of production (particularly in A1 areas) and marketing (mostly informal). This will require new, and better attuned, data collection techniques.
Unfortunately too often the emergency, humanitarian aid and disaster relief momentum overrides discussion of the developmental issues, and the scramble for food aid (and all the associated politicking) diverts attention and resources. As I have mentioned in this blog many times before, rural development challenges are many. They include the need to invest in irrigation to offset drought vulnerability, the importance of investment and reforms to ensure timely supply of inputs, a pricing and market policy to balance incentives between food and cash crops, a livestock policy that ensures such assets are secure and available in times of need, and, overall, more concerted support for the resettlement areas to ensure that they can indeed supply the nation with food.
Next week, I will continue this theme and look at the data on production and imports over time in a bit more detail. Since 2000 there is little doubt Zimbabwe is in a new era, and policy responses have to take this into account.