What have been the patterns of crop production in the Masvingo study areas in the past few years? The Masvingo data reflects the broader pattern nationally, with poor maize harvests over the past few years. National aggregate statistics (with all the health warnings that need to be attached) show maize production at 0.7m tonnes (2009), 1.2m t (2010), 1.5m t (2011), 1m t (2012) and 0.8m t (2013). This was consistently below the 1990s average of nearly 1.7m t; and mostly below the target of 1.6m t of maize or 2m t of cereals. Other crops have fared better, including sugar, cotton, and small grains.
However the overall picture has been mixed, and highly variable between farms and sites. While the aggregate picture has been dire, there are large variations, and some important anomalies that are significant when considering overall policy. In the resettlement areas of Masvingo for example, the level of production is significantly higher than assumed in national food security assessments for example. And for some groups – notably those in some of the A1 sites, significantly higher, with farmers producing several tonnes of maize and selling it.
In Masvingo, the 2009-10 to 2011-12 seasons were not good ones, with rainfall in Masvingo town 20-25% below the ten-year average of 675 mm. It was not just rainfall totals but also the pattern of rainfall that caused problems, with sudden downpours and extended droughts. This highly variable pattern has been a recurrent feature in recent years, and seems to be a pattern that is well established (and indeed recognised by the expert meteorologists, as well as farmers). This last two seasons have bucked the trend with 2012-13 rainfall 19% above the average (although poorly distributed) and this year has seen very high rainfall and the prospect of good harvests, although in some areas there was too much.
Across our A1/informal sites, then, maize production per household was on average just 779kg in the four seasons 2010-13. It was of course highly differentiated, with higher levels in the wetter areas of Gutu and Masvingo (with a very stark contrast between areas in 2012 and 2013), and overall output was highest in the higher ‘success groups’ (those generally with more assets). The A1 ‘self-contained’ farms had the highest production where 2.7, 2.0, 2.1 and 1.1 tonnes of maize per household were produced across these years. On the A2 farms there was huge variation in all years, with some producing practically nothing while a few have been regularly producing significant quantities across all years.
One indicator of ‘success’ we used before was the proportion of farms producing over a tonne of maize each year (the amount required to feed an ‘average’ family). While this was high in the good rainfall years of 2006 and 2008, the proportion of households in this category declined, with 32% producing more than a tonne of maize in 2010 and only 23% in the subsequent three harvest seasons. Again, the A1 self-contained farmers performed the best, with 56%, 53%, 56% and 39% of households producing over a tonne across the four years. However these figures are considerably higher than that seen more generally according to food security assessments. The Zimvac study for 2013-14 predicted that only 2% of households would have enough food from own production to provide needs. In this respect, the resettlement farms, even in dryland Masvingo, are faring much better and some are producing surpluses which will be important in supplying others in deficit.
Across the 2010-13 period 34%, 17%, 44% and 9% of A1/informal households sold some maize, although amounts were highly variable. For example, in 2010 in the self-contained A1 sites 45% of households sold maize, and 28% sold over a tonne, even in this poor season. In subsequent seasons 27%, 34% and 10% of A1 self-contained farmers sold over a tonne, showing a real commitment to commercial agriculture.
Yet across these years, there were still half to two-thirds of the sample who were struggling to produce enough and not selling any surplus, particularly in the drier areas of Chiredzi and Mwenezi. Although they also produced considerable quantities of sorghum and millet, it was not enough for many. This is of course disappointing, given the ambitions of the resettlement areas driving improved food security more broadly, and underlines the importance of access to water and irrigation.
Can it all be blamed on variable rainfall? Probably not. The level of production in the years following land clearance was boosted by the inherent soil fertility of the land. This has declined, and it has not been replaced by significant additions of manure and fertiliser. Total amounts of inorganic fertiliser applied have been low, with between 25% and 50% not applying any, and many applying very little. In the drier areas virtually none is applied. While cattle numbers are up, manure amounts are insufficient to cover the larger areas in resettlement farms.
In next week’s blog, I will look at the wider livelihood setting, putting relatively poor crop production in low rainfall years in the context of changes in on-farm assets, as well as off-farm income generating activities.
The on-going Masvingo study research is conducted by Ian Scoones, Blasio Mavedzenge, Felix Murimbarimba and Jacob Mahenehene.