Journal of Social Development in Africa (1990), 5,1,73-81 
Case Study 
What Are We Promoting? 
A Case Study of the Introduction of a New Milling 
Technology in a Rural Area in Zimbabwe 
BRIAN MACGARRY S J* 
ABSTRACT 
This case study looks at the introduction to Zimbabwe of a new sorghum milling technology 
developed in Botswana. This technology, it was assumed, would, by making the processing 
of small grains less laborious, encourage increased production of these grains. In the analysis 
of what actually happened in the milling project in one community focus is on the mill as a 
business enterprise and on issues of community organisation and management Thecase study 
concludes with the view that while there may be a number of non-monetary benefits accruing 
from the milling project, there has not been a noticeable increase in sorghum production nor 
is the financial self-sustainability of the mill clear. 
Introduction 
A project designed to introduce a new milling technology raises questions as to whose 
development is actually served, and the answers are not entirely simple. 
The story began in 1981. Due to personal contacts between the staff of SilveiraHouseand 
people who had developed and promoted a mechanical sorghum dehuller for installation in 
rural grinding mills in Botswana, Silveira House was asked if it would be interested in 
promoting this technology in Zimbabwe. The primary benefit of mechanical dehulling was 
that it greatly reduced the work involved in preparing meal from sorghum or millet grain. 
Without it, these grains are much more laborious to process than is maize, and this, it was 
assumed, was one reason why maize was becoming more popular than these small grains, even 
in drier regions where it could not give consistently adequate yields and the performance of 
the small grains is more reliable. TheintroductionofthedehullersinBotswanahad,inafew 
years, vastly reduced imports of maize (from South Africa) and produced a great increase in 
tbeconsumrXionofsc»^nummeal,llKHighttobed«tou^neweaseofniilUngattheviuage 
level. Zimbabwe was oneof several countries where, itwasfdt,Uwdehullertechnok)gy could 
produce similar benefits for people living in more arid regions. 
+ Reieudi Officer, Silv«anH«ue,PO Be* 545, Harare, Zimbabwe.
74 Brian MacGarry 
The backers of the programme, Canada's International Development Research Centre 
(IDRQ, envisaged the setting up of one or more mills, fitted with dehullers, in areas where 
sorghum or millet would be the most suitable crop for the climate. They hoped to observe: 
a) whether the availability of mechanical dehulling influences the preference of rural 
consumer/producers as between maize and the climatically more suitable sorghum 
b) the economic viability of the project, and 
c) social and technical factors likely to affect its acceptance in the arid regions of 
Zimbabwe. 
IDRC also hoped that the operators of the pilot mill/s would become active promoters of the 
spreadof technology inZimbabwe. They invited SilveiraHouse to propose sitesforpilot mills, 
and identify persons with sufficient local experience and technical and managerial skill to 
manage, run and maintain the mill, as well as to spread knowledge of the technology and its 
benefits, and possibly train intending operators of further such mills. A senior member of 
Silveira House's agriculture promotion staff identified a site at Sasa business centre, 25 km 
east of Mutoko in north-east Zimbabwe. He claimed that all the required conditions were met 
at this site. From the beginning, it later appeared, the project was viewed differently by the 
promoters and by the community at Sasa, the majority of whom seemed aware only that they 
had been offered a grinding mill and jobs for some of them running it 
Silveira House had had contact for some time with small cooperative groups of peasant 
farmers in Mutoko district, and paid a part-time promoter for these groups. There were six 
groups, all within 10km of Sasa, in 198 land the number of groups and the area they covered 
were both expanding rapidly. The mill was to be run by a cooperative formed by members 
of these groups. This cooperative would, it was hoped, purchase the mill building and 
equipment out of income from the mill, thus proving the viability of the project, transferring 
control of it to the local community, and making the original capital available to repeat the 
process elsewhere. 
Mill operations made a slow start, due to problems of both staffing and equipment for the 
project. It was thus May 1984 before the first day of grain milling took place on the site, and 
further technical problems delayed fulltime commercial operation till July of that year. 
The development of the project from this point can be considered under three aspects: 
1. the effect of introducing the dehuller on production and consumption of small grains 
around Sasa, and the spread of dehuller technology through the demonstration value 
of the mill 
2. the financial viability and consumer support for the mill as a source of meal (whether 
of maize, millet or sorghum) for domestic consumption, and 
3. the development of community organisation and self-management of the Sasa mill. 
Effect on sorghum/millet production and consumption 
Although the mill has operated more days per year than its competitors in the neighbourhood, 
and at lower prices, it has not reversed the decline in production and consumption of small 
grains. The traditional grain crop, and staple food, in this area was bulrush millet (mhunga), 
but by the time this project was proposed most farmers were growing maize as their staple food
Case Study. What Are We Promoting? T. 
and main cash crop, with enough finger millet (zviyo, rupoko) for home beer brewing, and 
very little mhunga. Those who were able to irrigate a garden or small field would grow some 
rice. Sorghum (mapfunde) was not grown, except for a sweet variety (magunde) grown for 
its sweet cane, not for its grain, in small gardens. 
The presence of the mill did not produce an immediate increase in the amount of mhunga 
being grown or consumed in the area, and did not even stem the decline in production in the 
first three years of operation. Table 1 summarises the data available for the first 4 years of 
operation. A similar picture emerges if we consider the weeks in which mhunga made up more 
than 20% of the total grain processed at the mill (Table 2). 
These figures show two tilings about mhunga consumption: 
a) It usually peaks in December, since sadza (the staple food, a thick porridge eaten with 
relish) made from mhunga is a preferred dish at festive times. 
b) Mhunga is still used as a food reserve to some extent. The 1984 grain crop was very 
poor, so that for most of the first three months of 1985 the mill only worked a four day 
week, instead of the usual six days, and in this time mhunga made up a largeproportion 
of the grain processed. 
If the incomplete figures for 1987-8 can be taken to indicate an increase, however slight, in 
the production and consumption of mhunga as compared to maize, the mill is unlikely to have 
influenced this directly. As can be seen, by this time the dehuller was hardly used at all. 
Reasons given by mill customers for not dehulling mhunga at the mill, even when they 
brought it to be ground, were: 
'double cost' of paying for grinding and again for dehulling (although the cost of 
grinding mhunga rose from 60c per bucket of 18 kg in 1984 to 80c in 1988, the price 
of dehulling remained 40c/bucket throughout). 
- machine dehulled mhunga meal did not have the slightly toasted taste of hand-dehulled 
mhunga, which is due to moistening the grain and later drying it over a fire. 
- (rarely) too much mhunga is lost in dehulling, (although the bran is given back to the 
customer and used as chicken feed). 
The perceived cost seemed to be the reason which carried most weight 
Reasons given for not planting more mhunga were: 
- mhunga suffers much more from the depredations of the quelea birds than does maize 
- if mhunga is planted in the same field for 2 or 3 years, it exhausts the soil 
- low producer price for mhunga. 
The last two reasons given show that, whereas maize came in as a cash crop and thus was 
cultivated from the start using artificial fertiliser and pesticides, the local mind-set still sees 
mhunga as a subsistence crop to be cultivated in the traditional way, with no purchased inputs, 
soyieldsarelowandtheysoondropwithdroppingsoilfertility. Incomeperhectareistherefore 
low, although the producer price per tonne has been higher than for maize since 1984. 
It is interesting to note that the first mapfunde (sorghum) milled at Sasa appeared in 
December 1988. Thismay bearesponsetotheproblemofcroplosstothebirds,sincesorghum, 
like millet, requires less rain than maize, but it is less attractive to birds because of the high 
tannin content in the grain hull.
76 £rian MacGarry 
Sasa mill as a business enterprise 
The optimistic picture of great demand for milling in the whole Mutoko district, painted by 
the Silveira House staff member who proposed that the mill be sited as Sasa, looked less likely 
when it was observed that Sasa had several business centres each about 5 km distant, and they 
already had grinding mills. A snap survey of sales of soap, salt, sugar, bread and green 
vegetables from the two stores at Sasa indicated that only SO households shopped there. This 
would mean approximately 230-260 people using Central Statistical Office (1984) data, or 
possibly300peopleusingourownsamplingofsizeof householdsusingthemill. This number 
would give a maximum grain demand in the expected catchment areas of 115-150 kg/day, or 
42-56 tonnes/year, clearly much less than the volume of business recorded in Table 1. 
The population estimate is possibly low, as would be the case if many households could 
not afford to buy the commodities sampled up to our estimate of their requirements, or if they 
bought significant proportion of them elsewhere, eg in towns. However, one day customer 
surveys have shown that since mid-1986, over 30% of the mill customers havecome from five 
km or more from the mill, and some of them from within one km of another mill. The Sasa 
mill is popular, and three reasons can be identified for this: 
- It is out of action due to breakdowns for less time than the other mills in the area. 
Breakdowns cost 24 working days each in 1984-85 and 1985-86,37 days in 1986-87, 
and 64 days in 1987-88. 
- Milling prices are lower than at other mills in the area. Table 3 lists prices for milling 
and dehulling per bucket (16,5kg maize, 18kg rapoko or mhunga) and compares them 
with the lowest price offered by competitors. 
- There is a sense in the community that 'this mill is ours', it belongs to a cooperative 
of which many members of the community are members, and it is perceived as trying 
to serve the whole community, and not operate solely for profit 
This last point will be elaborated in the next section, butfirst we must consider the commercial 
viability of the enterprise. 
The mill is able to employ two operators and pay them a regular wage, which has risen 
more or less in line with government directives, and to pay all running and maintenance costs. 
Silveira House, along with an installation team who came from Botswana, subsidised the 
setting up of the mill by providing skilled labour to direct the installation of the equipment 
(a more complex task than installing a simple hammer mill), by training the mill operators and 
by counting routine mechanical maintenance as part of its extension services for the first 
couple of years. Only the last of these items would distort the balance sheet if it were to be 
used as a guide to the viability of such a milling enterprise now that the technology is known 
in Zimbabwe. Even since the cooperative began paying all their own repair and maintenance 
costs, milling income has covered all operating costs. 
There remains the question of paying for replacement of equipment, or in this case, 
repayingtheinitialloan.outofincomc. Themillmachineryand someof the building materials 
represent a loan of $10 000 to the cooperative from Silveira House, of which $2 000 has 
been repaid to date (May 1989). Against this must be set the fact that cooperative members 
have contributed a total of $3 848 as share capital, and a sum of about $5 000 'disappeared'
Case Study. What AreWe Promoting? 77 
during 1986-87. The cause has been identified, and procedures tightened up to make 
misappropriation of cooperative funds much more difficult in future. A total profit of a few 
hundred dollars has apparently been banked since January 1988, so, had it not been for this 
incident, some$7 000-7 500)three-quartersoftheinitialcost,couldhavebeenrepaid,roughly 
half from income and half from share capital. 
A number of delays, beyond the control of either the cooperative or Silveira House, 
between the initiation of the project and its becoming operational, would have made any 
request for interest payment an undue burden on the cooperative members. Even now, the mill 
wouM have difficulty paying interest However the design of the building was more elaborate, 
and the dehuller bigger, than was necessary. The whole plan of the mill was standard for a 
large Tswana village (over 10 000 inhabitants) where grain throughout, and thus income, 
would be higher than at almost any rural centre in Zimbabwe. A smaller, simpler unit would 
serve the needs of a community like Sasa and, bringing in the same income for less capital 
expenditures and lower overheads, could be a financially attractive proposition. 
The fact that members are still prepared to contribute share capital with little hope of 
an early cash dividend shows mat they are still happy that their investmentbrings them benefits 
- an invisible cash benefit in die form of lower milling costs, and other benefits which may 
not be assessable at all in financial terms. 
Community organisation and self-management 
The delays mentioned above, affecting the delivery and installation of equipment, provided 
useful time in which to adjust the expectatkmsof all concerned in the project to what they could 
realistically expect out of it. A large part of the Sasa community apparently believed, even 
after the mill began operating (July 1984, after 3 years of talks, visits and discussions), mat 
they were merely being given a mill to grind their grain and provide jobs for some of their 
number. The ultimate promoters of the project, IDRC, on die other hand, had very high 
expectations of the levels of mechanical, managerial and promotional skills that might be 
found in such a community, and indeed, might be combined in one or two individuals. The 
Sasa mill was to be a pilot project, to test die economic and technical viability of diis dehuller 
technology under die conditions prevailing in Zimbabwe, to suggest possible modifications 
in die design of die equipment or operating methods, to propagate knowledge about the new 
technology available, and to be a centre for training further operators for similar mills. Two 
mill staff were to be employed. IDRC wanted diem to be bom technically skilled and 
experiencedas development extenskm workers, whileSilveiraHouse'sphilosophy emphasised 
that die two should be the choice of their community for the jobs. 
In fact, die two originally chosen, whom I shall call A and B, did bodi have experience as 
extension workers with Silveira House, one in agriculture and die other in youth work, but 
neither could claim much mechanical skill. It turned out that both had been chosen by the 
Silveira House staff member who had proposed siting die mill at Sasa, even diough odier staff 
members had doubts aboutB'sabihty to cope with areally responsible job. Shortly after they 
were selected, A died suddenly. In order to have two trainees to send for a course on using, 
maintaining and promoting die milling technology in Botswana, die original Silveira House
78 Brian MacGarry 
promoter chose C, a catechist for the local Catholic mission, to replace A. B and C duly 
attended the course in Botswana, where their trainers rapidly perceived that they would not 
be able to do much more than run the mill, and reshaped their training course accordingly. 
The equipment was supposed to be installed and the operators to begin work as soon as 
possible after their return, but then the delays began. 
In order to be able to drive either a hammer mill or the dehuller separately, or the two 
together, and to change between these possibilities without stopping the engine, the diesel 
engine needed to be fitted with a clutch. This is not required for the standard diesel engine 
and hammermill set used in rural grinding mills, nor for the other uses, such as water pumping, 
to which this type of engine is usually put in Zimbabwe. Separate arrangements, therefore, 
needed to be made to import a clutch. Finding one, ordering it, and arranging shipping and 
customs clearance, took about two years. This meant that when the trainee operators returned 
from Botswana in July 1982 there was no mill for them to operate, but also that the community 
had some timein which toconsiderwhether.andhowfar.theycouldmeetlDRC'sexpectations 
on monitoringtheprojectandpromotingits wider dissemination. At first thecommunity asked 
to have a manager appointed for the mill, but before one could be found popular support had 
swung behind operator B for the job. During 1983 the project marked time, a committee was 
chosen by the community to oversee the running of the mill, and it seemed advisable to pay 
a variable monthly sum to B and C out of project promotion funds as a retainer. The sum was 
intentionally variable from month to month. No contract had been entered into that would 
entitle B and C to a regular salary before there was work for them, nor oblige them to wait for 
that work, but this payment was meant to make it easier for them to wait. After about a year 
of this arrangement, B began demanding further payments, including a sum 'in lieu of leave'. 
This convinced the community of his unsuitability for the job, and, led by the new committee, 
they elected D, a local farmer, whose main qualifications for the job were a great capacity for 
hard work and sound common sense. C and D were thus the designated mill operators who 
welcomed an installation and training team from Botswana in May 1984. Further minor 
technical adjustments meant that commercial operation only started on 20th July 1984. 
At this time a manager still had not been found, so it was agreed that staff from Silveira 
House would help monitor the mill 'soperation. The mill operators agreed to record separately 
the daily income from grinding maize, mhunga and other grains (in practice, 'other' grains 
meant rupoko or chimera, a mixture of malted rupoko and maize for brewing) and from 
dehulling/n/jKHga,as wellasdaily diesel consumption. Occasional visits from theCentre were 
made to collect this data and to make some measurements to determine optimum operating 
conditions (retention time and size of load) for the dehuller. On such visits customer surveys 
were also sometimes carried out, to give the weight of each load of grain processed that day, 
the distance it was brought to the mill, the size of the customer's household and how often they 
came to the mill. 
A separate manager, with the time and skill to undertake wider promotion work for the 
dehuller, was never found. A possible candidate was employed for a short time, but proved 
unsuitable and noone else could be found locally. The community were unwilling to accept 
an outsider in this capacity.
Case Study. What Are We Promoting? 75 
Until 1987, although the mill income was being used to pay for fuel, operators' salaries 
and repairs, this was done in the accounts books of Silveira House, with the mill operators only 
recording income, and therefore not being as immediately aware of expenditure. This system 
had its weaknesses, as when the $5 000 disappeared, on its way from Sasa to Silveira House, 
between January 1986andMay 1987. Handlngoverfull control to themill committee reduced 
the chance of a repetition of the disappearance, by making all financial operations more open 
to inspection and checks by cooperative members. In early 1987, the cooperative opened a 
bank account, into which income was paid. In June they began paying the mill operators' 
salaries directly from this account, and from the end of July they began ordering their bulk 
diesel fuel and lubricating oil supplies and paying directly from their account. Finally, in 
September, they contracted with a mechanic, from some distance (due to lack of skills locally), 
for maintenance and repairs, whom they paid directly. 
By 1987 thecommunity finally grasped theimplicationsof the loan agreementunderwhich 
the mill had been set. They had requested from Silveira House, and received, training on 
cooperative principles and management They became more serious about collecting joining 
fees and share capital from cooperative members and this proceeded faster after a new 
committee had been elected to replace the one chosen in 1983. The 1987 election was 
accompanied by complaints about misuse of money, but, strangely, inadequate accounting for 
a few hundred dollars spent on food for an opening ceremony and lunch, and other occasions, 
from committee members at the mill caused more of a stir than the apparent direct theft of 
$5 000. It was only when the cooperative began payingtheir own bills directly that they started 
keeping records of expenditure, but once they did the rate of members' contributions to the 
cooperative picked up. By May 1989,14 members had each paid their full $200 share and 
another 13 had contributed a total of $1 048 as part shares varying from $20 to $140. The 
present committee has served for nearly two years, but there are indications that members will 
want some changes in 1989. 
In 1988, as they took full control of their financial management, the committee decided 
that, while recording all expenditure in the cashbook kept at the mill was essential, recording 
the income from processing each kind of grain separately was not, so it will not be possible 
to compare the quantities of different grains processed as fully as was done for 1984-87, but 
the mill will run well enough without thatinformation. Spot customer surveys, taking account 
of seasonal variations in mhunga supply and consumption, can indicate large changes in 
planting and consumption patterns, but the mill committee declined to take on the job of doing 
these surveys. 
Conclusion 
The introduction of the mechanical dehuller has not noticeably caused any swing back to 
growing and consuming mhunga in this community. Whether, even without any financial 
malpractice, this mill could be fully financially self-sustaining is still an open question, though 
altering theplan of the buildingandequipmentfor any similar mill in future wouldreduce costs 
without reducing income and thus improve the situation. The decisive factoratSasais the deep 
loyalty of the wholecommunity.even non-members of thecooperative,to their mill. An article
80 Brian MacGarry 
like this might have missed some of the non-monetary benefits that should be taken intc 
account beside the apparent financial loss, but the users have drawn up their balance sheet, am 
on theirs they are gaining from having the mill. 
From studying theoperarion of this mill, it seems most likely that thebestthatrural grinding 
mills usually can achieve is to cover running costs and simple maintenance and repairs, bo 
accumulating funds to eventuallyreplacetheequipmentisprobably beyond their capacity. N< 
examplecomes to mind that would contradict this statement On the other hand, thecontinirinj 
strong community support for the Sasa mill, and the knowledge that costs can be reduced ir 
future mills, has induced Silveira House to help at least one more community who asked fa 
assistance to set up a cooperative maize grinding mill to get a loan to start their business. 
References 
Central Statistical Office (1984) 1982 Population Census, A Preliminary Assessment. 
Harare. 
Zimbabwe National Household Survey Capability Programme (1985) Report on the 
Demographic Socio-Economic Survey, Communal Land of Mashonaland East 
Province 1983/4, CSO, Harare. 
Table 1 
Mhunga processed at Sasa Mill, 1/10/84 - 30/09/88 
absolute quantities and as proportion of all grains 
Dates Total Grain Mhunga 
processed Milled %of Dehulled %of 
tonnes tonnes total tonnes total 
6,5 279 
364 
5.0 
2,6 
1.8 
0,7 1
1
2
8 1/10/84-30/9/85 
1/10/85-30/9/86 
0,9 
3,3 
6,2 13 
>11 
0,4 
(negligible) 
211 
203 >5,4 
1/10/86-30/9/87 
1/10/87-30/9/88* 
for part of the year millers did not differentiate mA«/t£a from other grains in their records.
Table 2 
Weeks when mhunga accounted for more than 20% of business 
at Sasa Mill, October 1984 • December 1988* 
Week beginning 
17 December 1984 
31 December 1984 
7 January 1985 
14 January 
21 January 
28 January 
11 February 
18 February 
25 February 
18 March 
15 December 1986 
22 December 
2 May 1988 
19 December 1988 
Mhunga was not recorded separately from other grains processed January - June 1988, except the 
one week shown here, nor on Ac last three days (30% of the week's business) before Christmas 1988. 
Dates 
May'84 - 31/12/85 
1/1/86 - 17/10/87 
19/10/87 - 30/06/88 
- 2/07/88 
Case Study. What AreWe Promoting? 81 
Mhunga as % of total grain processed 
27,2 
28,4 
29,9 
243 
263 
30,4 
20,8 
22,6 
203 
20,0 
22,2 
29,1 
31,4* 
>20,0* 
Table 3 
MiUing charges at Sasa 
Dehulling Hammer Milling/Bucket Nearest 
mhunga mhunga rupoko maize competitor 
/bkt maize/bkt 
50c 
60c 
45c 
50c 
60c 
65c 
45c 
50c 
40c 
40c 
80c 
90c 
70c 
80c 
60c 
75c 
60c 
75c 
40c 
40c