Contents - Next

This is the old United Nations University website. Visit the new site at

Editorial policy


The Food and Nutrition Bulletin is intended to make available policy analyses, state-of-the-art summaries, and original scientific articles relating to multidisciplinary efforts to alleviate the problems of hunger and malnutrition in the developing world. It is not intended for the publication of scientific articles of principal interest only to individuals in a single discipline or within a single country or region. Notices of relevant books and other publications will be published if they are received for review. The Bulletin is also a vehicle for notices of forthcoming international meetings that satisfy the above criteria and for summaries of such meetings.

The Food and Nutrition Bulletin also serves as the principal outlet for the publication of reports of working groups and other activities of the UN ACC Sub-committee on Nutrition (SCN) and its Advisory Group on Nutrition. The SCN itself is a focal point for coordinating activities of FAO, WHO, UNICEF, the UNU, Unesco, the World Bank, the World Food Programme, the World Food Council, the United Nations Environment Programme, and other bodies of the United Nations system which have an interest in food and nutrition.

Unsolicited manuscripts of articles of the type published in this and previous issues may be sent to the editor at the Cambridge office address given above. They must be typed, double-spaced, with complete references and must include original copy for any figures used (see the " Note for contributors" in the back of this issue). All articles submitted will be reviewed promptly and the author will be notified of the editorial decision. Any disciplinary or conceptual approach relevant to problems of world hunger and malnutrition is welcome, and controversy over some of the articles is anticipated. Letters to the editor are encouraged and will be printed if judged to have an adequate basis and to be of sufficient general interest.

It is expressly understood that articles published in the Bulletin do not necessarily represent the views of the United Nations University or of any United Nations organization. The views expressed and the accuracy of the information on which they are based are the responsibility of the authors. Some articles in the Bulletin are reports of various international committees and working groups and do represent the consensus of the individuals involved; whether or not they also represent the opinions or policies of the sponsoring organizations is expressly stated.

Nutritional anthropology and agriculture

Health and nutrition effects of sugar cane production in south-western Kenya

Eileen Kennedy and Ruth Oniang'o


In 1984, a two-part study was initiated to evaluate the income and nutritional effects of shifting from maize to sugar cane production in south Nyanza, Kenya. During the first phase, baseline data was collected on the socioeconomic, food-consumption, and health and nutritional status of a cohort of households prior to their entry into the small-holder sugar cane out-growers scheme. The combination of the two studies provided one of the few opportunities to have baseline economic and health information on the households prior to their entry into cash cropping. In the first phase, the per capita income of the two groups was the same. In the follow-up of the cohort sample, the per capita income of the new entrant group was higher than that of the non-sugar group. This increase in income did not appear to influence pre-schoolers' morbidity or growth. Data from both phases of the study indicate that the health and sanitation environment had the most impact on pre-schoolers' growth, suggesting that growth will not be substantially improved in the short term by income alone.


The commercialization of agriculture is the cornerstone of economic development in many developing countries. Proponents of strategies advocating an emphasis on commercial crops (often called cash crops) see this as a means of generating and saving foreign exchange, increasing the incomes of the rural small holder, providing employment for the landless, and stimulating growth linkages with other segments of the economy.

Critics of the acceleration of the production of export and cash crops have argued that not only have the economic benefits not materialized, but in some cases the transition to commercial agriculture has had a negative effect on staple food production and hence household-level food security as well as health and nutritional status. Many of the most contentious nutrition issues in the debate have revolved around the impact of commercial agriculture on women and pre-schoolers.

In 1984, at the request of the government of Kenya, a study was initiated to evaluate the income and nutritional effects of shifting from maize to sugar cane production. The government was concerned that in areas undergoing this transition, particularly the production of sugar cane, household-level food security and pre-schooler nutritional status were deteriorating.

Conceptual framework and research approach

Almost all of the previous research on the nutritional effects of cash crop production has concentrated on evaluating outcomes [1]. Is cash cropping good or bad? This approach is simplistic, since presumably cash crops can have different impacts on income, consumption, and health. More important, the results of these types of studies-whether positive or negative-have limited usefulness for policy formulation. Emphasis exclusively on outcomes tells us nothing about the mechanisms through which commercial agriculture affects health and nutrition.

The purposes of the study discussed here were twofold: to evaluate the impact of the shift from semi subsistence to commercial agriculture on income, expenditures, food consumption, nutritional status, and health; and to identify the process leading to these outcomes.

Figure 1. Some linkages between cash-crop production and nutritional status at the household and intra-household levels

provides a conceptual model of the pathways through which cash crop production can potentially influence the health and nutritional status of individuals within the household. Past research concentrated on a limited number of household-level variables, mainly agricultural production. Noticeably absent is any research related to an assessment of the effects of cash cropping on intra-household dynamics.

The schema presented in figure 1 served as the basis for conceptualizing and designing the study. The first phase took place from June 1984 to March 1985. This was followed by a second phase from December 1985 to October 1987. The sample included 504 representative households from the community; this was one of the few studies on commercial agriculture to include agricultural as well as non-agricultural households from the community served by the South Nyanza Sugar Scheme (SONY).

The 1984-1985 study was able to collect baseline data on the socio-economic, food-consumption, and health and nutritional status of a cohort of households prior to their entry into the small-holder sugar cane outgrowers scheme. This group has been designated as the new entrants. They were followed until the time of payment for the first harvest.

Most prior research on the topic of commercialization of agriculture has involved ex-post studies only. Households were evaluated only after they had already entered a cash cropping scheme. Accordingly, it has been difficult to ascertain the baseline health and nutritional status of households and individuals prior to their making the transition. The combination of the two studies provided one of the few opportunities to have baseline economic and health information available prior to the entry of households into cash cropping.


Income and food consumption

Proponents of a strategy advocating the commercialization of agriculture have assumed that farmers' incomes would increase as they switched all or part of their land to cash crop production. While higher income is only one of a series of household objectives, it is clearly an important one.

The results from the 1984-1985 study indicated that the incomes of sugar cane farmers who received at least one payment for the crop were significantly higher than those of the non-sugar cane producers (table 1). Sugar cane accounted for the major portion of the differential between the groups.

TABLE 1. Comparison of annual income per capita for the cohort group (Kenyan shillings)


Mean nominal income

Mean real incomea



New entrants







Sugar farmers with income






















Wage earners














Sample mean







source Ref 2.
Numbers in parentheses equal number of households.
a. 1985-1986 incomes adjusted to 1984 levels using GDP deflator [3,4]
b. sugar farmers significantly (P<0.05) higher income than non-sugar and landless groups.
c. New entrants versus non-sugar farmers (P < 0.05)
d. Sugar farmers versus non-sugar farmers (P < 0.05).
e. Merchant households significantly (P < 0.05) higher income than landless, non-sugar farmer and sugar farmer groups.

The income of the new entrant group (1,956 shillings per capita) was virtually identical to that of the non-sugar cane producers (KSh 1,924 per capita) in the 1984-1985 study. It is noteworthy, therefore, that in the follow-up study, the incomes per capita (both nominal and real) of the same new entrant group were significantly higher than those of the non-sugar group. In the follow-up study, both the new entrants and the sugar farmers had per capita incomes that were significantly higher than the non-sugar cane producers.

Part of this difference in incomes was due to differences in marketed agricultural income, KSh 791 and KSh 365 respectively. Other sources of income also contributed to the difference. Of the KSh 1,129 difference in nominal income per capita between the new entrants and the non-sugar cohort, 41% was contributed by commercial agricultural income, 38"/o by semi-subsistence income, and the remaining 21% by higher non-farm incomes in the new entrant group.

Income is one of the major determinants of household energy intake (table 2). Total household income has a positive and significant effect on household caloric intake, but one relationship is non-linear. At the mean caloric intake, the income elasticity of demand for calories is 0.15.

TABLE 2. Household consumption function


T statistic

Semi-subsistence Y (%)



round 2



round 3



round 4



Head of household schooling



Adult equivalent






Income squared



Non-farm income (%)



Relocated dummy






Dependent variable equals total daily household caloric intake.

Interestingly, different sources of income have different effects on household energy intake, and these effects are above and beyond the pure income effect. The percentage of non-farm income has a significant and negative effect on caloric consumption. In addition, there is a beneficial effect from having income from production used for home consumption. This favours the non-sugar farmers, of whose income 49% comes from semi-subsistence production.

One major reason for the different effects of various sources of income on energy consumption may relate to control of income within the household. Non-farm income is earned and controlled primarily by men. Much of the agricultural production used for home consumption comes from female-controlled plots of land, and is more likely than male-controlled income to contribute to household energy.

Female-controlled income and income from agricultural production are more likely than other forms of income to be used to enhance household food security [5]. The data from the present study also suggest that semi-subsistence income has more of a positive effect on energy intake than other forms of income. It may be that, in addition to the issue of control of income (male versus female), the real or perceived transaction costs of converting food crop income into cash may make it more likely that semi subsistence production contributes to household food security.

Energy consumption has a negative and significant association with years of schooling of the head of household. Males who are more educated are generally wealthier in this community, as is typical in most countries. It appears that increased wealth is related to a more varied diet rather than simply a more energy-dense diet. Similar results are seen when total expenditures are used as a proxy for income.


Morbidity patterns of pre-schoolers

The sugar cane scheme is one form of development assistance that was directed towards the south Nyanza district with the expectation that the economic growth generated by the outgrowers' programme would result in an improved health and nutritional status for the population, and in particular for the vulnerable groups-pre-school-age children and pregnant and lactating women. The study allows us to assess whether income changes in the longer term are associated with decreases in morbidity.

Table 3 presents data on morbidity patterns for women and pre-schoolers for all four rounds combined. For the cohort sample, there was no significant difference in the total time ill or the time ill with diarrhoea for pre-schoolers across any of the activity groups. The significant gains in income for the new entrant group have not translated into a decrease in the average time ill for pre-schoolers.

TABLE 3. Time ill and time ill with diarrhoea for preschoolers and women, cohort sample




Total time ill (%)

Time ill with diarrhoea (%)

Total time ill (%)

New entrant

29.5 (85)


24.5 (32)


29.8 (428)


23.8 (168)


31.2 (542)


24.3 (220)


20.8 (45)


21.8 (16)

Wage earner

31.6 (30)


31.9 (14)


31.6 (62)


21.8 (31)

Sample X

30.3 (1,192)


24.1 (481)

Source: Ref. 2.
Based on average of all rounds.
No two groups significantly different.
Numbers in parentheses equal number of women or children.

Morbidity patterns for women and pre-schoolers were analysed within per capita quartiles. For both groups, there were no significant differences across income quartiles in the total percentage of time ill (table 4). There also were no differences for pre-schoolers across income quartiles in the total time ill with diarrhoea.

TABLE 4. Time ill and time ill with diarrhoea for preschoolers and women stratified by per capita income quartiles







Pre-schoolers total time ill (%)









time ill with diarrhoea (%)









Women total time ill (%)









Source: Ref 2. Numbers in parentheses equal number of women or children.
No significant difference among groups.


Nutritional status of pro-schoolers

The present research was conducted in an area of Kenya with the highest mortality rate from birth to age two years-216 per 1,000-of any part of the country. The area also has a high prevalence of pre-schooler malnutrition. The government hopes that one positive effect of the transition from semi subsistence to commercial agriculture will be improvements in general well-being, including in child health and mortality rates.

TABLE 5. Z scores for children in studies 1 and 2


Study 1 Z score

Study 2 Z score (all-round average)







New entrants - 1.46 (90) - 1.13 (90) -0.27 (90) - 1.74 (61) - 1.06 (61) 0.005 (61)
Sugar farmers - 1.34 (356) - 1.03 (356) -0.22 (356) - 1.67 (243) - 1.14 (243) -0.15 (241)
Non-sugar farmers - 1.50 (556) - 1.17 (556) -0.31 (556) - 1.76 (349) - 1.10 (353) -0.04 (349)
Merchants - 0.99 (62) - 0.86 (62) - 0.27 (62) - 1.05 (29) - 0.89 (29) - 0.26 (29)
Wage earners - 1.65 (30) - 1.49 (30) -0.59 (30) - 1.87 (24) - 1.49 (24) - 0.51 (24)
Landless - 1.45 (77) - 1.06 (77) -0.18 (77) - 1.99 (40) - 1.36 (40) -0.16 (39)
Sample mean - 1.42 (1,171) - 1 11 (1,171) -0.28 (1,171) - 1.72 (746) - 1.13 (749) -0.10 (743)

No two groups significantly different.
Numbers in parentheses equal number of subjects.

The Z scores for height for age, weight for age, and weight for height averaged for all four rounds for studies 1 and 2 are shown in table 5 (see TABLE 5. Z scores for children in studies 1 and 2). These data are on the cohort of children who were in both studies. This sample was therefore older in study 2. In the 19841985 study, there were no significant differences in any of the three anthropometric indicators across any of the three groups. This was somewhat surprising, given that in study 1 the incomes of the sugar farmers were approximately 25% higher that those of the non sugar farmers and new entrant groups. In the later study, however, the same results emerged; no significant differences were found across any of the households in any of the three indicators, despite the fact that the new entrants had an average income per capita that was significantly higher than the non-sugar group.

TABLE 6. Regressions of pre-schoolers' Z scores for height/age, weight/age, and weight/height. 1985-1987




Weight, height

Independent variable


T statistic


T statistic


T statistic

Head of household -0.134 -0.74 -0.137 -0.99 -0.03 -0.20
Household size -0.012 -1.51 -4.28-03 -0.70 1.44-03 0.216
Sex (1 = boy) 0.090 1.09 0.03 0.40 -0.03 -0.47
Female head of household -0.211 -1.45 -0.14 -1.29 -0.07 -0.54
Average Z score
study 1 0.605 20.67 0.55 20.09 0.39 10.64
Child's calories 2.42-04 1.96 2.02-04 2.13 1.10-04 1.07
% diarrhoea- -0.02 -3.36 -0.014 -3.03 -7.09-03 - 1.46
Mother's height 0.026 3.69 7.80-03 1.46 - 6.3-03 1.09
Area (in ha) 2.98-03 0.29 0.01 1.36 0.01 1.29
Age 0.025 8.10 0.01 6.21 1.27-03 0.51
Constant -5.60 -4.56 -2.03 -2.16    




Table 6 (see TABLE 6. Regressions of pre-schoolers' Z scores for height/age, weight/age, and weight/height. 1985-1987) presents the Z score regressions for height for age, weight for age, and weight for height; these multivariate analyses substantiated much of what appeared in the descriptive analyses. One interesting finding for all three Z scores is the strong relationship between the baseline score from the earlier study (1984) and the anthropometric indicator from the follow-up study (1985-1987). Given that the time span between the baseline period and end of the later study is two and a half years, one might not expect this effect to be so strong. For height, an increase in the height-for-age Z score of 1.0 in the baseline period was associated with a 0.61 increase in the current height-forage Z score. The corresponding values for weight for age and weight for height were 0.55 and 0.39 respectively. This would indicate that children who were doing well in the 1984-1985 study had a high probability of doing well in the follow-up period. Conversely, those who were not doing well earlier had a high probability of continuing to have less than optimum nutritional status. These data give credence to those who advocate faltering growth as a major criterion for identifying children at risk of long-term growth problems.

Pre-schoolers' caloric intake is a significant, positive determinant of height and weight but not weight for height. Similarly, the percentage of time ill with diarrhoea has a negative effect on height and weight but not weight for height. Finally, in general, as children grow older, their Z scores for height for age and weight for age improve.


The government of Kenya has been interested in the appropriate balance between cash crop and food crop production for domestic consumption. There has been some concern that in areas with increased cash cropping, particularly increased sugar cane production, pre-schooler nutritional status has deteriorated.

The present study is one of the few based on a random sample of farmers in a commercial outgrowers' scheme. In addition, it is one of the few studies to date that provides a community assessment of the range of effects of commercial agriculture. One premise was that some of the most dramatic effects of cash cropping might be on households not directly in volved in the scheme-the landless and the merchants.

The study design was methodologically much stronger than is usual for research of this type; the new entrant group on whom baseline information was available prior to their entry into the sugar cane outgrowers' scheme could be followed until the time of first harvest. This allowed stronger inferences to be made about the actual impact of commercial agriculture.

The results suggest some positive effects of commercial agriculture on household income. In the 1984-1985 study, the incomes per capita of the two groups were virtually identical. In the follow-up of the cohort sample, the income per capita of the new entrant group was KSh 1,129 higher than the non-sugar group. Part of this increment was due to differences in commercial agriculture income. The new entrants earned KSh 791 per capita from commercial agriculture compared to KSh 365 per capita for the non-sugar producers. Here again, in 1984, the commercial agriculture income per capita of the new entrants (KSh 404) and the non-sugar farmers (KSh 393) was almost identical. Entry into sugar cane production thus increased incomes.

The sugar cane outgrowers' programme as it is implemented in Kenya was associated with a significant increment in income. This resulted in positive effects on the household energy consumption of the new entrant group. This benefit, however, did not appear to have influenced pre-schoolers' morbidity or growth. There is a growing awareness that family-level factors may be poor indicators of a child's nutritional status.

Many governments and international agencies are putting increased emphasis on income-generating schemes as a way of achieving health and nutrition objectives. While increases in income may be a necessary condition, by themselves they may not be sufficient to alleviate malnutrition, at least in the short term.

Data from both studies suggest that it was the health and sanitation environment that had the most impact on preschoolers' growth. One of the major determinants of child growth was the growth pattern, which was influenced significantly by pre-schoolers' morbidity patterns and the health and sanitation environment. Children who were not doing well earlier continued to record inadequate growth. This suggests that without improvements in factors that influence their health, pre-schoolers' growth will not be substantially improved in the short term by income alone. More emphasis must be placed on the health implications of agricultural policies and projects, with particular attention paid to ways to improve the health infrastructure in a given community.

The insignificant effect of income on health may relate to who within the household controls it. Sugar income is not seen as household income but rather as men's income (table 7). Not only is men's income different from women's but the expenditure responsibilities differ. It is therefore not surprising that the money earned from sugar production is spent on items like housing and school fees, categories of non-food expenditures that fall within the responsibility of men [6].

TABLE 7. Decision-making for food expenditures and sugar cane income


Households (%)

For food expenditures






other household members


For sugar cane income






other household members


do not know


Part of the difference in expenditures for men and women may also relate to the periodicity of income. Women's income in the form of food crops and trading activities comes in smaller, more continual amounts. This may influence how the money is spent. Men's income from sugar comes only after 18 to 24 months of work and is paid in one lump sum. Income from lump sums tends to be spent differently than that from small, regular sources.

Income is influenced by control (male versus female) and its pattern of flow into the household. If policymakers are interested in maximizing the nutritional effect of increased income, several steps might be taken. First, if it is culturally appropriate, the contract for the cash crop-in this case, sugar cane- should be in the name of both the head of the household and the wife (wives). This would help foster the perception of household income rather than simply male income.

Second, more periodic payment for the sugar cane crop-advances against anticipated production- might ensure that the marginal propensity to spend on food and other nurturing categories would be higher than it is with lump sums.

Finally, one issue that was not touched upon directly in the study warrants discussion. The community in which the outgrowers' scheme was implemented is one where malnutrition is endemic. Households may not be aware of the problem since their children look like all the others in the community. The outgrowers' programme involves approximately 30% of the households in the community and would be an excellent and visible way to reach a significant portion of the population regarding the nutritional needs of women and children. Nutrition education integrated into a primary health care delivery system could have a significant effect on the health and sanitary environment of children.

To date, most of the farmers who have joined the outgrowers' scheme have remained in it. It is naive to think that, given the way the programme now operates, there will be a mass exodus back to food crop production. Some fine tuning of the programme will help maximize the potential impact of the increased income on household and preschoolers' nutritional status.


  1. von Braun J, Kennedy E. Commercialization of subsistence agriculture: income and nutritional effects in developing countries. Working paper on commercialization of agriculture and nutrition no. 1. Washington, DC: International Food Policy Research Institute, 1986.
  2. Kennedy E. The effects of sugarcane production on food security, health, and nutrition in Kenya: a longitudinal analysis. Research Report no. 78. Washington, DC: International Food Policy Research Institute, 1989.
  3. World Bank. World development report, 1986. Oxford: Oxford University Press, 19X6.
  4. World Bank. World development report, 1987. Oxford: Oxford University Press, 1987.
  5. Greer J, Thorbecke E. Food, poverty, and consumption patterns in Kenya. Geneva: International Labour Organisation, 1986.
  6. Kennedy E, Cogill B. Income and nutritional effects of the commercialization of agriculture in southwestern Kenya. Research Report no. 63. Washington, DC: International Food Policy Research Institute. 1987.

Contents - Next