Improvement to storage on the farm

Contents - Previous - Next

The case for improvements in storage

As indicated previously, storage involves substantial costs and risks as well as potential benefits for farmers. Storage competes with other activities valued by farm family members, and it is necessary to understand where storage fits in to the entire farming system and household economy in order to assess the need for interventions and the probability of their uptake.

Over the past two decades the need for economic and social analysis in the planning and design of storage interventions has become more widely recognized. This stems from the realization that any 'improvements' in storage will only be attractive to farmers, traders or governments if the perceived benefits substantially outweigh the costs. Technical superiority is generally insufficient (although it can be attractive for its prestige value), and farmers and traders are likely to tolerate quite high storage losses before undertaking complex or expensive changes to their storage systems. An understanding of the reasons why people store, and the systems within which storage occurs, is necessary in order to estimate how the benefits and costs of innovations are likely to be assessed by the intended users of the technology.

Rates of adoption of new storage technologies at the farm level have often been disappointing (Phillips 1981; Compton, 1992). In some cases, projects have failed because they were promoted on the basis of assumptions which turned out to be false. Sometimes it was incorrectly assumed that storage ranked high among farmers' lists of priorities. From such experiences it can be concluded that, before storage projects are implemented, there is a general need for more research into the economic and social factors involved.

Table 1.1. Comparison of costs for new permanent storage facilities, using different storage types, for long-term storage of grain (cost US$ per tonne of wheat)

It is also now generally accepted that local, established storage systems are usually well adapted to local conditions, and losses from grain storage are already low and acceptable to farmers (Greeley 1987, Compton 1992). This is not to say that improvements cannot be made. Indeed, the following factors point to an increased need for improvements in the handling and storage of grain at various levels in the system.

(i) Increasing urban demand

Due to demographic changes urban population in most developing countries is growing at 5 % or more per annum. In addition many countries, particularly those in Asia, are experiencing massive increases in intensive animal production, creating large markets for feed grains. For example, Indian poultry production grew by about 9% per annum in the 1980s. Consequently an increasing proportion of grain production is destined for the market rather than subsistence use, increasing storage requirements on the farm and elsewhere in the marketing chain.

(ii) Changes in government policy

Structural adjustment programmes and market liberalization in a number of African and Asian countries are increasing the role of the private sector in storing produce which is surplus to subsistence requirements. It was noted previously that in most African countries, traders and millers are not heavily engaged in storage, and this means that farmers in surplus producing areas are having to greatly increase their storage activity.

(iii) Changes in the farming system

On-farm and off-farm storage systems have been affected by technical change in other aspects of the farming system. The green revolution has involved the adoption of new varieties which are often more susceptible to storage losses (Golob and Muwalo, 1984). It has proved difficult for plant breeders to combine higher yields with storage durability, since the very qualities which lead to higher yields, and therefore (potentially) increased income also make the grain more attractive to pests. Thus, high yielding varieties of maize tend to give large, soft grains with less husk cover than traditional varieties.

Short duration varieties have allowed for increased cropping intensity. This can give rise to further storage problems when one of the harvests occurs in the wet season, making it difficult for farmers to dry the grain sufficiently for storage. Farmers in some areas have responded to the situation by growing high yielding varieties for immediate sale, and traditional varieties for storage and on-farm consumption (Giga and Katarere, 1986).

High yields also imply that farmers may need to manage the storage or sale of larger quantities of grain within a shorter space of time, which in itself may cause problems and encourage farmers to sell at harvest, in order to free up the labour for field preparation of the next crop. In some cases labour constraints at harvest lead to early or late harvesting of the crop, with consequent losses (Compton, 1992).

(iv) Changes in the pest population

A major change in the incidence of pests can prompt farmers to seek new storage technologies. In Tanzania the larger grain borer, a destructive pest of stored maize and cassava, was introduced from its native habitat in Central America in the early 1980s. Farmers were reported to have suffered up to 30 per cent losses from the new pest. In response to their demands the government, with donor assistance, implemented a successful extension programme to control the pest.

The larger grain borer has now spread to a number of other East and West African countries, including Kenya, Togo and Ghana, but losses in these countries have not yet reached the levels recorded in Tanzania.


What factors must one consider in assessing the potential for on-farm improvements?

The first step in the identification of appropriate technology is the assessment of the needs of potential users. In the case of post-harvest technology, claims of high losses and of the potential for reducing them have provided a major justification for the promotion of new technologies. The issue is discussed below in the next sub-section.

Whether or not there are good quality data on losses, it is also important to investigate the potential demand for the technology by its intended users. Even if losses appear quite high, it may be that post-harvest problems do not rank high among farmers' priorities. It may also be that they are more concerned to reduce labour or other costs than they are to reduce losses. Mechanical threshers and mills have been widely adopted in Bangladesh even though they tend to increase losses, because of the savings in labour costs (Greeley, 1987). As a result women labourers from poor households have lost a source of income from hand threshing and milling.

Even where there is a demand for loss reducing technical changes, farmers may find it difficult to adopt recommended technologies, because of cash flow problems, labour constraints, lack of materials, or storage chemicals. Small farmers and traders often find it difficult to obtain credit at reasonable interest rates, since formal financial institutions consider loans to them be too risky.

If it is decided that some form of intervention is both desirable and feasible, then the full range of options should be considered. For example, if storage losses are high, then, in addition to investigating storage technologies, the potential for altering cultivation and postharvest activities (e.g. shelling maize instead of storing it on the cob), for introducing varieties with improved storage characteristics, or for experimenting with biological control methods can also be examined. A discussion of a wide range of options is given in Compton (1992).

Notwithstanding these options, the most successful storage technology to date appears to be the use of insecticides. They can easily be integrated into existing storage systems, and often give high returns. The main constraints on increased insecticide use are: availability of appropriate insecticides at the right time; stability of the formulations used; farmer training in the correct types and correct use of insecticide; cost, which sometimes renders their usage uneconomical.

In view of the latter, it may be appropriate to use locally available materials, such as woodash, sand or certain plant materials which control the growth of insect populations. Use of such materials is most likely to be viable where small quantities of grain are involved, for example in storing locally produced seeds. When farmers have to store larger amounts of grain (e.g. a tonne), usage of such materials may prove tedious and cumbersome, and sufficient quantities of them may not be available. At the same time some of these materials may have toxicological effects which have yet to be investigated. Research in the coming years should throw more light on the usability of a range of these materials.

Introducing new store types has often proved difficult. The main reason is that the capital cost of new stores is too high, and often fails to offset the reduction in the value of losses, especially where stores are not used to full capacity. There have been notable successes in the introduction of metal bins in Swaziland, Central America and the Punjab area of India and neighbouring Pakistan, but no such cases are known of in poorer areas of Africa. For similar reasons, mechanical driers have also been difficult to promote. Unless there are severe drying problems, sun drying tends to be preferred since it is cheaper. The improved quality of mechanically dried grain is rarely reflected in a higher price, and therefore provides no incentives for farmer adoption.

As well as assessing the level of losses and the demand for the new technology, one must also appraise the cost-benefit or financial viability of the improvement to the individual farmer. The next three sub-sections discuss the assessment of the three factors highlighted above, i.e. losses, demand and financial viability.

(i) Assessment of storage losses

Losses can occur at several stages of the post-harvest chain, including threshing, storage, transport, milling, wholesale and retail distribution. In order to decide whether it is worth taking action over losses of any sort, one should obtain information on losses at all these stages.

There has been a tendency to overestimate storage losses, and to base estimates on extreme cases or guess-work rather than on sound empirical testing. Figures of 30 per cent or more are not uncommon (Greeley, 1987, p. 13ff). By contrast, the results of detailed field studies suggest that under traditional storage systems in tropical countries losses are typically around 5 per cent over a storage season (Tyler and Boxall, 1984), depending upon the crop, the ambient conditions, the period of storage and other factors. Somewhat higher levels have been encountered in the wetter parts of West Africa and Central America.

Loss figures around the 5% level should not however be considered insignificant. Firstly it should be noted that physical losses are usually accompanied by qualitative losses affecting the mass of the grain in store. Secondly the losses are mainly experienced during the lean season before the new harvest is ripe, thereby having an adverse effect on the food security of farming families at a particularly critical period. In Honduras, farmers' feelings of insecurity about this period have been an important motive for adopting metal storage bins.

Even where detailed studies are undertaken, there are a number of methodological difficulties involved in estimating losses (Greeley, 1991). Loss assessment methods tend to be slow and to require skilled field and laboratory staff. They are often undertaken on experimental sites, making it difficult to relate the results to on-farm situations.

There are a number of factors which tend to lead to an upward bias in the loss estimates. Firstly, extremes may be taken rather than averages. Ideally the sample size and standard deviation should be quoted with the loss estimate to avoid this. Secondly, removals from store over the season are not always accounted for. Where removals do occur, percentage losses calculated on the basis of grain remaining in store will be overestimates. Another source of overestimates lies in treating partial damage as a total loss, when in fact the damaged grain would be used by farmers for home consumption or animal feed. A fourth source of upward bias lies in the potential for double counting losses at different stages in the post-harvest system. Losses at one level are related to those at other levels.

Another difficulty in using estimates of losses to justify technical change is the problem of assigning to the losses a value which makes sense to the potential user of the technology. The most common form in which losses are expressed is as a percentage weight loss. But what is important from the farmer's point of view is the use that the grain can be put to, or the market price that will be received. Grain intended for sale may be consumed, or that intended for consumption used as animal feed.

A rapid loss assessment method for estimating storage losses in maize and cassava has recently been developed in Togo (Compton et al. 1992). The method attempts to incorporate farmer criteria in defining categories of loss, and since the measurement occurs in the field, rather than at a laboratory, results can be discussed with farmers on the spot. Such methods could usefully be integrated into post-harvest technology projects.

(ii) Assessment of demand

As in all market research one should start with desk research, including the interviewing of key informants, to gain whatever data are readily available about storage systems, the uptake of improvements introduced in the past and other relevant information.

One should then visit representative villages in the area of interest to analyse the farming system together with the farmers (i.e. carry out a participatory rural appraisal or PRA) to identify opportunities for improvements, taking care to interview a representative sample of farmers and womenfolk, including significant minority groups likely to have an important role in crop storage (e.g. larger mechanised farmers). Out of this activity should come: (a) an assessment of whether any improvements are worth considering in greater depth; (b) a list and description of these ideas or concepts suited to particular groups of farmers.

Selection of technologies may be aided by means of matrix analysis. This involves tabulating the alternative technologies (on a horizontal axis) against the full range of criteria used in their selection (on a vertical axis). Each technology can be scored or ranked in terms of the respondents' perception of its performance against each criterion. The different steps to be considered in matrix analysis are outlined in the shaded box on the next page, using the appraisal of alternative storage structures as an example. Typical results are displayed in Figure 1.4.

The same approach could be used in assessing pest control systems, or other technologies.

Figure 1.4. Matrix Scoring for three grain storage structures

  Metal bin Improved crib Traditional crib
Durability of structure ***** ***** **
Ease of handling ***** ***** ****
  * *  
Peace of mind ***** *** ***
Low construction costs * *** *****
Does not attract pests ***** **** **
Total Score: 27 21 21



  1. Find a group of key informants who are knowledgeable and willing to discuss. Hold an open-ended discussion with them about on-farm storage. This should be along the lines of a 'focus-group interview', with the moderator guiding the discussion in a non-directional manner, with occasional prompts, and asking for clarification of points of interest.
  2. As the discussion proceeds, ask the participants to consider a range of alternative storage structures. Ask them which are of interest, and the pros and cons are for each. Probe for further criteria.
  3. Based on the discussion, make a short list of (a) storage structures worthy of further consideration, and (b) the most important criteria by which the informants appraise their suitability. Then construct a matrix with the storage structures displayed on the horizontal axis, and the criteria on the vertical axis. This can be done on a large sheet of paper, or on the ground (using seeds, stones etc. for scoring).
  4. Make negative criteria positive (e.g. 'attracts pests' is written on the chart as 'does not attract pests').
  5. Convene a group of villagers to assess each storage structure following the list of criteria. Structures may be scored on a scale of, say, ten points. If many items have to be compared, a smaller scale with less points can be used. The highest total score will indicate the most preferred.
  6. Ask which criteria are most important? If it emerges that some criteria are far more important for the villagers than others then give them more weight by multiplying their scores by a weight factor (e.g. 2).

Adapted from: J. Mascarenhas, Participatory rural appraisal and participatory learning methods, recent experience from MYRADA and South India; Forests, Trees and People Newsletter No. 15/16.

This exercise can be carried out with people representing different social groups.


In this case the metal bin has the highest score. However, if 'low construction costs' were of leading importance to the villagers, resulting in the multiplication of its scores by two, then the traditional bin would come out top, with 31 points, compared to 28 and 24 respectively for the other two constructions.

As an alternative to scoring, data may be ranked against each criterion, 1 = best, 2 = next best, etc.. Ranking can be used for up to 7 items. The method is straightforward and rapidly elicits much information on why participants give priority to certain criteria. However ranked data for different criteria cannot be added up. Ranking only conveys an order of preference, but not the degree of liking or disliking.

By this stage it may be concluded that certain technologies are affordable and can be field tested without further research. Where this is not the case however, or where one is approaching a number of villages with different characteristics, one may proceed to appraise the different options through concept testing. This market research technique was introduced by NRI into Swiss-funded storage projects in Central America, and is now used for the rapid appraisal of storage concepts, without going to the expense of constructing prototypes.

In a concept test, respondents (chosen from the target population) are shown a picture or mock-up of the new storage structure with a list of their key attributes, and then answer questions about likes and dislikes, how and when he/she would use the store, willingness to invest in one, preference between alternative concepts and so on. The tests can be carried out in individual depth interviews or in group interviews, or in a combination of both.

The tests can yield some quantitative information e.g. about the percentage of respondents wishing to invest in the structure, about their ranking of technologies etc., but above all it produces in a short space of time a lot of qualitative insights into farmers' thinking about storage and its place in the farming system and the household economy.

(iii) Assessment of financial viability in on-farm storage and handling

On-farm improvements offer the potential for an increase in net benefits through a reduction in variable costs, such as labour, a reduction in the value of grain losses, or an increase in the market value of the grain as a result of using the technology.

An example of the use of cost-benefit analysis (CBA) on farm-storage projects is that undertaken by Boxall and Bickersteth (1991). They compared seven different storage technologies in terms of the break-even price which the farmer must obtain on one bag of maize to cover storage costs, at two different discount rates. Their findings are presented in Table 1.2, and details of the methodology used are shown in Annex 1.

They found that traditional systems with lower capital costs and no operating costs achieve lower break-even prices in spite of higher losses. Various development programmes had favoured the improved storage crib, but this technology proved the least favourable on account of its high capital costs. Only in areas with particularly high losses would the improved crib be financially viable. The mud bin was the most cost-effective structure because of its durability, cheap construction cost and low losses.

The conclusion reached was that storage technologies currently being recommended under certain development programmes would not, under normal circumstances, be profitable for farmers. Similar findings were encountered in two other West African studies (Al Hassan, 1989; Stabrawa, 1992).


Marketing of new on-farm technology

Having established that a given concept is desired and financially viable, one can then proceed to test-market the technology, by persuading farmers in a given locality to build or install prototypes. As with any other marketing exercises, the test-market will need to be supported by a delivery structure (involving artisans, trainers etc.), after-sales service and maybe credit. The demand for the technology should be monitored through the growth of sales in the target area, and the level of market penetration i.e. the number of units installed/potential number installed. Reasons for non-adoption should be analysed, with a view to either changing the product or the marketing strategy, or revising downward one's view of market potential. Figure 5 (see Figure 1.5. Schematic presentation of Technology Dissemination) summarises NRI's approach to technology assessment and dissemination in Guatemala.

Table 1 2. GHANA - Maize STORAGE COSTS ( In Cedis; Cedis 360 = US$ 1)

Contents - Previous - Next