Tag Archives: fewsnet

Food security in Zimbabwe: why a more sophisticated response is needed


The food security situation in Zimbabwe – and indeed across large swathes of southern Africa – is serious. El Niño has struck hard and production levels this past season were well down. The UN estimates that in Zimbabwe alone 4.1 million people – 42% of the rural population – will be in need of support before the next season. Aid agencies are raising funds and are involved in a major humanitarian operation (see WFP and USAID, for example).

We are now entering the most difficult period. Between September and March, when early ‘green’ crops become available, the food situation will be tough, and many will be reliant on handouts and purchased imported food. Disposal of livelihood assets is already occurring and FEWSNET predicts that large parts of southern Zimbabwe will be in ‘emergency’ conditions, together with parts of Mozambique and Malawi.

There is little doubt that the harvests this year were really poor. And this was on the back of a bad season last year. This means that stocks are low and funds circulating in the local, rural economy limited. I do not want to question for a minute the severity of the situation, but I do want to challenge the way it is being portrayed, and ask whether this allows for the most effective targeting of those really in need.

Data challenges

For Zimbabwe the basic data comes from the annual ZimVac report, complemented by various crop surveys. ZimVac, as discussed on this blog before, is a major survey based on a sample of 14,434 rural households across 60 districts. Enumeration areas are chosen across districts and samples selected based  on population density estimates from the most recent population census. It assesses food production, cash income, livestock and so on, and comes up with a food access estimate, based on a daily 2100 k Calorie intake requirement during the consumption year to 31 March. Those unable to meet food needs through a range of sources are deemed to be in deficit and in need of support. This is where the 4.1 million figure comes from – the number of people estimated to be in this situation at the end of March 2017 (even if just for a day).

But as discussed before on this blog, these estimates may miss out on certain aspects. For example, In April, when visiting field sites in some areas hit badly by drought, I was surprised how much maize was being produced in home gardens and around settlements this year. While the main field crop had failed, more intensive production near the home. Sometimes involving supplementary irrigation, and certainly higher inputs of organic fertiliser, home garden areas were producing maize, including substantial quantities of green mealies. These crops rarely get noticed in the larger censuses as they focus on the main field crop, but added up these can be significant, although of course totals are way down on other years.

The other missing story relates to livestock. This year there were major concerns that the El Niño drought would decimate livestock. There were significant die-offs early on, but thankfully sporadic rains fell in February. This was too late for most crops, but it did replenish grass and water sources in many parts of the country, including those drought prone areas of Masvingo and Matabeleland that were suffering livestock mortalities. This turn-around will have had major impacts on food provisioning in these areas in the absence of harvests. There were entrepreneurs buying up animals in numbers and this was a ready source of cash for many. Many livestock were moved to resettlement areas where there is more plentiful grass due to (currently) lower population densities. The high livestock populations in resettlement areas, particularly in southern districts, adds to their food security resilience.

Livestock and their movement is often forgotten in food security assessments (ZimVac covers elements of this, but it’s complex, and difficult to capture in large surveys). Along with the importance of green mealies, other ‘famine’ crops, and the range of (often illegal) coping strategies that people employ mean that successful food provisioning is far more extensive than the UN agencies suggest.

While the data is broken down by district, it is not differentiated by the type land tenure and use. We do not get a sense of the differential vulnerabilities of, for example, communal area dwellers, those with A1 or A2 farms, villagised or self-contained, nor workers linked to such rural households. We know from extensive research that rural communities are highly differentiated, both within and between sites. At the moment we get a very blunt assessment, district by district. The report lists the ten best-off and worse-off districts, for example. Some of the districts where we work, where there was more land redistribution, both in the Highveld and further south, are in the better-off areas. Does this mean land reform areas are less food insecure? We cannot tell from ZimVac data as presented.

A more complex pattern: why land reform is not to blame

There are hints though that a more complex pattern sits below the aggregate numbers. The ZimVac summary report (p. 150) shows that nationally only 11% of households will be food secure this year based on their own cereal crop production. This is even lower in drought-prone areas, such as Masvingo, for example. On aggregate 58% of the national rural population will be food secure through the consumption season, but this is made up through access to income from a variety of sources, not just food production. How do these aggregate figures match up with data from the new resettlement areas?

We’ve been tracking food production in our study areas in Masvingo for some years. In our sites in Masvingo and Gutu districts for example across the harvest seasons from 2003 to 2013, between 44% and 69% of households produced enough for household consumption (estimated at 1 MT). In the Wondezo extension A1 site in Masvingo, farmers produced on average 2 MT in 2014 and over 6 MT in 2015, with 85% and 89% producing sufficient from maize alone for household consumption in those years. In our A1 resettlement sites in Mazowe, over 5 years between 2010 and 2014 seasons the average household maize production was 3.5 MT, declining over time as tobacco production increased. This means that on average 78% of households produced more than a tonne of maize in each year, and were food secure from own-farm production alone. This of course does not account for the significant cash income from tobacco in Mazowe (realising nearly $3000 per household on average across A1 farms between 2010 and 14), or vegetable production and livestock in Masvingo, along with other sources of income.

In other words, the ZimVac sample must be very different. 11 per cent this year (and higher but still low figures in other years) having sufficient food from own production is way lower than in our admittedly much smaller samples in the resettlements. In our areas, consistently over time and across sites, we do not see the level of food insecurity recorded by the ZimVac surveys – although of course it exists in pockets, among certain vulnerable people. There are of course communal areas nearby our A1 sites where the situation is quite different, and it is probably from here that the ZimVac data derives. Our comparisons with communal areas showed the contrasts, with resettlement areas outperforming communal areas across the board. But without any differentiated national food security data, it is difficult to make sense of the aggregates generated by standard crop assessments and livelihood surveys.

This food security crisis therefore is not the result of land reform as some would have it (as I keep telling journalists who ask; here’s an example from a Dutch daily that offered a more sophisticated take). Other countries in the region have suffered badly from the same drought, and Zimbabwe has before, long before the post 2000 land reform. In fact, land reform areas are an important part of why the actual underlying situation is better than it might be. My hunch – still not tested despite much encouragement – is that ZimVac’s sampling frame (appropriately for a national sample that is proportional to population density) is focused on communal areas. This means that the dynamics of the new resettlements in the food economy are being missed out on.

As reported many times on this blog, we see significant flows of food and other finance coming from the A1 resettlement areas, both to communal areas and to urban centres, through kin networks and labour migrancy. This is unrecorded and therefore not accounted for. My guess is that it is really significant in the overall food security story in the country, and taking account of land reform in the wider assessment would allow a redirection of effort by humanitarian and development agencies to support production for boosting local food security and economies, investing where the potential lies.

There is no reason for complacency though. Things could and should be much better, with proper investment. For example, the lack of irrigation infrastructure (and its state of repair, and its poor functioning due to intermittent electricity supplies) is a cause for major concern, and undermines resilience

The politics of food aid: why a more targeted approach is needed

Food aid is of course is highly political. It always has been, and accusations of partisan allocations have occurred again this year. Many are happy not to rely on the obligations and patronage that food aid implies – whether to the party-state or NGOs – and seek their own way. But there are some who are really destitute, without the networks that provide support. They are really needy and include a lot of people, but it’s certainly not 4.1 million. They include widows or older parents without living children, child-headed households, farm labourers, those with illness and disability, for example.

They all need help, as existing provisioning and coping strategies are insufficient. They are scattered all across the country – including in the high potential, richer areas within communities who are otherwise prospering, and are difficult to find. These are the people who need food, and would be a better focus for a more sophisticated, targeted approach to relief, which could combine with a more strategic developmental approach to increase production and market led economic development across communal, resettlement and urban areas.

This post was written by Ian Scoones and appeared on Zimbabweland

Leave a comment

Filed under Uncategorized

Millions at risk of food insecurity in Zimbabwe? Or not? How the dire predictions were confounded by a good harvest

Last September I critiqued the assumptions behind the prediction that 2.2 million people would be needing food aid. In order to raise funds and galvanise attention, international agencies, local lobby groups and the media were using an extreme worst case scenario figure, based on a variety of assumptions, many of them highly questionable.

As it turned out, the rains arrived and a good season has followed (with some exceptions of course). In the section below, I offer some extracts from the most recent USAID-funded FEWSNET update on the food security situation in Zimbabwe. Good rains have boosted production and the current food security projections to September are largely very positive.

It is amazing what a change in the weather can do. But it also adds to my earlier plea to be cautious about headline figures and assumptions in forward projections. There is no harm in being cautious – this must be the sensible stance – but overblown figures and dramatic proclamations that serve particular interests should be guarded against.

Unlike the portrayals of imminent doom, the relatively good news about a reasonable harvest does not hit the headlines, or raise aid money, and the bad news stories from Zimbabwe persist. So for a change, and in case you are not regular readers of FEWSNET bulletins, I thought you would like an update on a good harvest and a reasonably positive food security situation

Here is a summary edited from from the May update:

The majority of very poor households across the country including the traditionally food insecure southwestern districts, will experience Minimal (IPC Phase 1) acute food insecurity outcomes between May and June owing to the projected above average 2013/14 harvest. Similar outcomes will continue from July through September as most households will still be consuming cereals from own production.

Markets will continue functioning but most of the cereal supplies are likely to be locally procured with a few imports by private traders. As households begin to access cereal from their own production there have been significant reductions in monthly maize grain price trends. Since March, national maize grain prices have dropped by 11 percent, but in comparison to national averages during the same period last year the prices are still 16 percent higher. For maize meal the national average stands at $0.66 and has decreased by 2 percent in comparison to the same time the previous month, but remains 4 percent higher than the national average for same time last year. Month-on-month maize grain prices fell by 26 and 16 percent in Manicaland and Masvingo Provinces, respectively.

Casual labor opportunities are projected to increase by up to 20 percent throughout the outlook period as a result of ongoing harvesting activities. Additional incomes, particularly in the northern areas, will be earned through tobacco preparation, sales and casual labor for poor households. However given cash constraints, most casual labor will likely be paid by in-kind.

The first round results of the Ministry’s crop and livestock assessment indicate that there are increased chances of an above average harvest, especially for maize, millet, and sorghum. This assumption is based on an estimated 16 percent increase in cropped area for cereals this season in comparison to the 2012/13 season. Maize alone this season accounts for approximately 1.6 million hectares, which is an 18 percent increase from the previous season. This increase in area planted for cereals is due to fairly well distributed rainfall patterns this season.

Ongoing tobacco curing and sales are boosting household income, particularly in the northern areas, where production levels are projected to have significantly increased. Based on the first round assessment, this year’s production levels has surpassed the 2012-13 season by about 21 percent. At the household level, higher than average tobacco production will increase farmer income levels and opportunities for casual labor opportunities (i.e. curing, processing, transportation) for poor households. Households benefiting from this labor will therefore receive additional income for food purchases and other livelihood needs.

Cotton production this season is 16 percent below last year’s levels. The processing of cotton is ongoing in cotton growing areas but incomes are likely to remain low. The reduction in the area under cotton is due to marketing price uncertainty given the low marketing prices offered during the previous season.

The increase in the availability of water due to the good rainfall this season will increase gardening activities from May through September. Vegetable production will provide both food and cash to very poor households.

Livestock body conditions in areas including Matebeleland South and Masvingo Provinces have significantly improved and are in good shape. Despite the improved pasture and water access for cattle, the calving rate included in the recent first round crop and livestock assessment report remains low at 49 percent, and only 2 percent higher than last season.


The FEWSNET report provides the assumptions it uses in this analysis, along with some useful graphics. The second assessment report is due shortly and this will update the situation. Certainly the tobacco harvest looks promising, and reports from many parts of the country shows grain production is good.

So, thankfully 2.2 million people in Zimbabwe didn’t need food relief assistance, and the agricultural production has prospered in a good season. This however should be no reason for complacency. Droughts strike hard in a system where irrigation is not widespread, and improving resilience to such shocks must be a key part of future investments.

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


Filed under Uncategorized

Why good numbers matter in Zimbabwe (part II)

This week’s blog follows on directly from last week, when I introduced the excellent new book, Poor Numbers, by Morten Jerven. This week we move from the general argument to the Zimbabwe case.

Let me offer three examples – each of which have been mentioned in this blog before – that complement Jerven’s cases, and contribute to the same bigger point that good numbers matter.

Agricultural output data: Zimbabwe’s agricultural data comes from a variety of sources, including annual crop surveys, market surveys and assessments of throughput at marketing depots. In the past, when the sector was dominated by a few large farms, it was relatively easy to get a picture of production each year. Output from the communal areas was assessed through state marketing channels through marketing boards for most of the agricultural commodities, especially maize (but also cotton, tobacco and beef). While statistics on cotton and tobacco remain reasonably good, as their marketing is channelled through few players, the production and marketing of maize and beef, by contrast, has changed dramatically since land reform.

Today there are diverse marketing channels, including much locally-focused marketing and little reliance on the old marketing board routes. And with many more farms across the country (around 150,000 new units in the A1 schemes alone), field-level monitoring by extension agents is nigh on impossible. For important crops such as the small grains (millets, sorghum), groundnuts, many oilseeds and beans, as well as smallstock, we know virtually nothing about total production and marketing.

The bottom line is that we don’t know how much food is produced and where, nor do we know how much is stored and marketed. Despite the attempts of Fewsnet, ZimVAC and others, the estimates are increasingly guesswork, especially as sampling frames and data collection protocols have not changed sufficiently to respond to the dramatically reconfigured agrarian structure.

Each year we get conflicting estimates of how dire the harvest is going to be, and the consequences this will have for food imports, and food aid. With such uncertainties, this becomes a critical area of political contestation: between government and the donors, and even between international agencies. Claiming a food ‘crisis’ may be the only way of securing international funds, as sustaining an ‘emergency’ has been essential to continued international engagement through ‘humanitarian’ aid. Such a response may well be justified; but it may be not. The problem is often we don’t know.

Migration data: Similar uncertainties centre population data and migration-related demography. While we know that migration, particularly to South Africa, has increased, we have absolutely no idea how many people have moved permanently there (or indeed to other destination countries, although the data for the UK, for example, is better). Large numbers are bandied around, which serve particular politically purposes; in South Africa (linked to xenophobic, anti-immigrant rhetoric) and in Zimbabwe and internationally (supporting the narrative that people are ‘fleeing’).

But the figures of course don’t take into account the long-term pattern of circular migration whereby people move temporarily, or indeed increasingly seasonally. If we were to believe the figures, there would be far fewer people in Zimbabwe than there seem to be. For example, the preliminary results for the 2012 census show that the population has increased by 1% over a decade and stands at nearly 13m. Even within the country we don’t know where people are living. There is an assumption that the urban areas are growing, as people flood to the cities. But is this the case? Debbie Potts doubts this data for sub-Saharan Africa generally, but until we get better locational census data that accounts for regular movement, we will not know.

Land ownership data: This is perhaps the most contested, and in the absence of a proper land audit, we cannot know. But when ‘surveys’ purport to present data that show that “40% of the land was seized by Mugabe and his cronies”, and these figures get reported in the international media as fact, we are in trouble. This most recent examples of this short-cut journalism and recycling of ‘facts’ are from the BBC (on the Hard Talk show with Patrick Chinamasa) and the UK Guardian (in a link put in by the paper in an otherwise good piece by Simukai Tinhu). The earlier land audits by Utete and Boka have shown categorically the problem of elite capture in the A2 sites, and our detailed province-specific work in Masvingo supports this. But the scale is nothing like that claimed.

This poverty of data leads to a poverty of understanding, and so a distortion of debate. We should not be ignoring the abuse of the land reform programme by some politically-military connected elites, and the ownership of multiple farms is clearly contrary to any regulation, but our focus should equally not be only on this issue, and the wider picture, based on realistic data, needs to be central. This is why, in terms of the GPA and in line with the now agreed constitutional commitments, a proper land ownership and use survey (an audit) is critical.

If you don’t know how much food is being produced, how many people are in the country or have left and who owns what land, then how can you begin to make plans for the future? As contributors to other headline statistics, including GDP, such figures may result in major distortions.

For example, in Zimbabwe, GDP figures have been used to show the dramatic decline, and then impressive recovery in the formal economy (see the shower of graphs in the most recent budget statement), yet, as I have argued before, even in the depths of the crisis in the late 2000s, economic activity was far higher than measured. The ‘real economy’ – informal, often based on barter exchanges, sometimes illegal, much of linked to cross-border trade – was thriving, despite the collapse in the core, formal economy. It had to: this is how people survived. If you believed the figures on the formal economy, where the numbers were collected, people would have been suffering far more than they did.

As the formal economy has recovered, this has been registered in the statistics, but the informal economy still exists, and indeed the 2000s saw a massive restructuring of economic activity, not only in the agricultural sector, but across the economy towards more small-scale, informally-based enterprises. This is not a bad thing, as it provides the basis for more inclusive, employment generating, broad based growth. But if it is not understood, measured and recorded, it does not feature in planning and crucially budget allocation discussions. ZIMSTAT has recently published the 2011/12 Poverty, Income Consumption and Expenditure survey, and in a future blog I will review its findings, and the degree to which it has been able to respond to the changed post-2000 context.

While it may seem that a focus on statistical services is a rather dry and dull subject, it is in fact essential. ZIMSTAT has a small ‘did you know?’ box on their website’s front page. It says: “The likely success of development policies in achieving their aims will be improved by the use of statistics”. They are right. Revitalising statistical services, and improving their capacity to carry out national-level, macro-census type work, as well as smaller, more focused surveys, complemented with qualitative insights, is vital.

If development is to be successful, a thorough-going and honest debate on the quality of data and how to improve it is essential. Jerven’s superb book discusses an important topic with clarity and honesty; and for donors thinking of investing in government capacities in Zimbabwe again, it is well worth a read.

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


Filed under Uncategorized