Contents - Previous - Next

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

The consequences of seasonal food insecurity for individual food-consumption patterns in north-western Benin

M. J. van Liere. E-A. D. Ategbo, A. P. Den Hartog, and J. G. A. J. Hautvast


Qualitative and quantitative dietary patterns were studied as indicators of seasonal changes in household food security in north-western Benin. These studies were carried out at both the individual level (women) and the intra-household level (husbands, wives, children 2-5 years old). Dietary changes between the intermediate period (April) and the pro-harvest period (August) included increased consumption of pulses and tubers, which compensated for a decrease in cereal consumption, and increased consumption of gifts, purchased foods, and wild foods as distinguished from foods from the household's own production. The adequacy of children's energy and protein intakes increased in the pre-harvest season, whereas those of the parents decreased Children and adults had same patterns with regard to the contribution of different food groups, except that children ate more pulses and less tubers than their parents in the pre-harvest season.


In many developing countries, especially in Africa, farmers depend mainly on their own production for their food supply. In unimodal climates this means that crops grown in one short rainy season yield the yearly supply of food. As a consequence, food availability is not constant throughout the year, and farmers often experience a food-insecure period just before the new harvest. Seasonal fluctuations in food availability and household responses to this insecurity can influence individual consumption patterns.

In studies of household food insecurity, qualitative and quantitative dietary changes are useful indicators [1]. Qualitative changes look at shifts in the use of different food sources such as wild or unconventional foods instead of those produced at home. Quantitative changes consider actual contributions to energy and protein intake. Shifts from preferred items to those of lower status and unconventional foods are normal in areas of food deficit, but may also indicate anticipated stress [2].

Most studies that report on seasonal adaptations of food patterns present data at the household level, whether qualitative or quantitative [3-10], and only some provide individual quantitative data on seasonal adaptations in consumption patterns [11, 12]. Seasonal reductions in food availability may not influence all members of the households equally. Findings from studies concerned with intro-household food distribution are not conclusive. Some cross-sectional studies have found an advantageous food allocation for adult males or boys over females [11,13, 14]; some have found no difference in nutrient adequacy for different age groups [15]. Some studies have found that children are protected from seasonal food shortages [12,16]; others have not [17].

To analyse how seasonal food insecurity and responses to it influence individual food-consumption patterns, we studied seasonal changes in the contributions of different food groups and food sources to the energy and protein intake of women, as well as in intra-household food distribution among husbands, wives, and children 2-5 years old.

Study area

The study was conducted in the commune of Manta in north-western Benin. A unimodal climate ensures seasonal fluctuations in food availability. Annual rainfall is 1,300 mm, concentrated between May and September. The region is part of the Guinean-Sudanian savannah with scattered trees. Tree crops with an economical or nutritional value are provided by the baobab (Adansonia digitata), shea tree (Butyrospermum parkii), African locust (Parkia biglobosa), and mango (Mangifera indica).

The main ethnic group, the Otammari, are chiefly subsistence farmers. The main food crops are sorghum (Sorgho spp.), millet (Pennisetum spp.), and fonio (Digitata exilis), but beans (Vigna spp.), bambara groundnuts (Voanzeia subterranea), and rice (Oryza sativa) are cultivated as well. Groundnuts (Arachis hypogea) are cultivated only as a cash crop. Home garden products include, among others, okra (Hibiscus esculentus), red peppers, and pumpkins.

Crops are dried and stored in mud granaries. At the end of the rainy season (August) stocks may become depleted; this "hungry" pro-harvest season ends with the harvest of fonio in September and October. Early varieties of maize, millet, and beans may be harvested in small quantities in July and August. The other crops are harvested between October and December.

January to April, the season with limited agricultural activities, is marked by many festivities and relatively high food availability. In August 1990 and 1991 respectively, 15% and 18% of households had cereal stocks of 24 x 10³ kcal (which equals about 10 days' energy requirement for one adult) or lower. In November 1991 the percentage of households with less than 24 x 10³ kcal in stock increased to 41% because of the failure of the fonio harvest and the depletion of the old cereal stocks.

Producing staple foods is the work of both men and women, but women take care of processing, preparing, and distributing foods as well as gathering wild foods. The average daily diet is two or three meals consisting of a thick porridge (pāte) of sorghum or millet flour served with a relish of green leafy vegetables or okra and condiments. Eating groups are formed, with individuals sharing one calabash of pate and one bowl of relish. Women eat with their younger children, and men eat together with their older sons.


Materials and methods


The subjects for the studies reported here were subgroups of a larger sample of 214 households selected for a longitudinal body-weight study, described elsewhere [18], from the total population of five villages (about 4,000 inhabitants) in the commune of Manta on the basis of criteria of participation in subsistence farming and permanent residence in the study area.

Women's food consumption

From the 214 households, 45 women were randomly selected to participate in the food-consumption study. The selection criteria were being 18-45 years old, non-pregnant, and non-lactating. Data were collected on four occasions: pro-harvest season (July-August) 1990, harvest season (mid-November-December) 1990, intermediate season (March-April) 1991, and pro-harvest season 1991. A full data set for the four seasons was obtained for 27 of these women. Reasons for being discontinued were pregnancy and low compliance with the food-source records in the first season.

Intra-household food distribution

In addition, an intro-household food distribution study was carried out on 26 households, also randomly selected from the 214 households. The characteristics of these households were monogamy and having at least one child between two and five years of age. Food-consumption was recorded during the intermediate season (March-April) 1991, when food was still sufficiently available, and the pro-harvest season (July-August) 1991, when food stocks were becoming depleted. One household dropped out; a full data set was obtained for 25 households.


Women's food consumption

The observed weighed record method [19] was used to measure food-consumption on four consecutive days, reflecting the local market cycle. All raw ingredients, prepared dishes, portions served to the women, and left-avers were weighed. The women were asked to serve individual portions for themselves instead of sharing the portion of their eating group. An assistant followed each woman during the day, carrying weighing scales with her in case the woman ate anything when she left the house. Scales with a capacity up to 2 kg were used for ingredients, portions, and left-avers, and scales up to 10 kg for prepared dishes. The source of each ingredient-i.e., whether it was provided by her own household's production (stock or field), purchased, a gift, or came from wild resources-was recorded.

The ingredients were classified after recording in food groups: cereals and tubers, meat and fish, pulses, vegetables and fruits, fats and oils, nuts, beer (mentioned separately because of presumed high intake), locust-pod flour (mentioned separately because it was expected to be important in the hungry season), and condiments. Tubers were not treated as a separate food group because data from another sub-study of the same research programme showed that they are consumed by only 7% of the households during the harvest season and rarely during other seasons. The energy and protein contributions of different food groups and food sources to total intake were calculated using food-composition tables [20, 21]. (The protein quality of the habitual diet was estimated at 85% digestibility. The adequacy of protein intake was therefore determined as 85% of the calculated protein intake divided by the protein requirement.)

Intra-household food distribution

Only meals prepared and consumed at home were considered for the intra-household food-distribution study. For these meals all raw ingredients and prepared dishes were measured. The women were asked to serve each participant his or her individual portion instead of sharing the portion with their eating group. The portions served to the husband, wife, and children were weighed, as were leftovers. Snacks or foods prepared and consumed outside the household were not considered, since the foci of this study were distribution patterns at the household level and the contribution of household resources. Energy and protein intake as well as the contributions of different food groups to the measured intake were calculated with the help of food-composition tables. Tubers were considered as a separate group here.

Data analysis and statistics

All data were analysed by means of SPSS-PC 4.0 software [22].

Women's food consumption

Food intake was averaged over the four measurement days. For five women, an average of three measurement days was used for the first period because the records for one of the days were not valid due to mistakes on the forms. The relative contributions of certain food groups and food sources to total energy and protein intake were compared between the seasons using Friedman two-way analysis of variance for non-parametric data [23]. When the chi-square value was significant, the Wilcoxon signed test was performed for the concerned variable to determine which periods differed significantly.

Intra-household food distribution

The energy and protein intakes provided by the principal meals eaten at home were expressed as a ratio of the individual requirements [24]. Requirements were calculated according to WHO/FAD recommendations [24]. The basal metabolic rate (BMR) was calculated with equations using body weight, and energy requirements were calculated as 2.10 x BMR for men and 1.82 x BMR for women in the pre-harvest season and 1.78 x BMR for men and 1.64 x BMR for women in the intermediate season. The protein needs of adults were estimated as 0.75 g/kg/day. A correction was made for lactating women (there were no pregnant women in this sub-group) of +480 kcal per day and +13 g of protein per day.

For the children, no individual data on body weight were available, but average body weight was calculated using the data set of the longitudinal anthropometry study [18] for the two periods, and energy requirements and safe levels of protein intake were calculated accordingly (table 1).

The adequacy of fulfilment of requirements was compared between seasons using Student's paired t-test. For each season, requirement fulfilments were compared between groups using one-way analysis of variance. Seasonal changes in requirement fulfilment among men, women, and children were also compared by one-way analysis of variance.

TABLE 1. Average body weights, energy requirements, and safe protein intake levels calculated for children in the intra-household food-distribution study, April and August 1991-Manta, Benin

Age (years) Weight (kg) Energy requirement (kcal/kg/day) Safe protein level (g/kg/day)
Apr Aug
2-3 11.2 12.0 103 1.13
3-4 13.1 13.7 97 1.09
4-5 14.2 14.7 94 1.06



Women's food consumption

The women's total energy intake ranged between 2,604 and 3,010 kcal, and their total protein intake between 73 and 84 g (table 2). The contributions of the different food groups are shown in table 3. Cereals and tubers were the mainstays of the diet, providing 52%-65% of the total energy intake and 46%-57% of the total protein intake; their highest contributions were during the harvest season and the lowest during the pre-harvest season. Pulses made the second highest contributions, supplying 11%-20% of energy and 18%-34% of protein intake, with the higher values in the pre-harvest season. Fats and oils contributed only 5%-10% to total energy intake, about the same as provided by local sorghum beer. The contributions of vegetables and fruits, meat or fish, and condiments were negligible. Locust-pod flour, a tree crop from Parkia biglobosa, is eaten in periods of food shortage, but its contribution was also negligible (1%).

Of the different sources of foods, the highest proportion came from the household's own production, that is, stock and field contributions (table 4). About 53%-65% of energy and 55%-70% of protein intake was provided by own stocks, and another 2%-11% of energy intake came directly from the field. The contributions of own production were highest in the harvest and intermediate periods. Foods gathered in the bush provided 6%-9% of energy intake but only 1%-2% of protein intake. Foods received as gifts made their highest contribution in the pre-harvest season (15%-20% of energy, 14%-21% of protein) and a somewhat lower contribution in the other two seasons (13% of energy, 12% of protein). Purchased foods were more important in the 1991 seasons (providing 11%-12% of energy and 12%-14% of protein) than in the 1990 seasons (5%-7% and 8%-9% respectively for energy and protein).

TABLE 2. Mean total daily energy and protein intakes of 27 women in four seasons-Manta, Benin

Season Energy (kcal) Protein (g)
Mean SD Mean SD
Pre-harvest, Aug '90 2,744 536 82 21
Harvest, Dec '90 3,010 723 84 27
Intermediate, Apr '91 2,604 576 73 20
Pre-harvest, Aug '91 2,633 577 77 19

TABLE 3. Relative contribution (percentages) of different food groups to the total energy and protein intakes shown in table 2

Food group Energy Protein
Aug Dec Apr Aug Aug Dec Apr Aug
'90 '90 '91 '91 '90 '90 '91 '91
Cereals/tubers 52 65a 62a 60a 46 56a 57a 53a
Meat/fish 2 1 2 2 6 5 7 7
Pulses 20 16a 11a 15c 34 27a 18a 28c
Vegetables/fruits 1 1 7a,b 1a,c 2 2 5a,b 2a,c
Fat 10 6a 5a 10b,c - - - -
Nuts 1 2 1 - 1 3 1 1b,c
Beer 10 6 10 8 8 5 8 6
Locust-pod flour 1 - - 1 - - - -
Condiments 2 2 2 2 3 3 3 4

a. Different from August 1990, p < .05.
b. Different from December 1990, p < .05.
c. Different from April 1991, p < .05.

Intra-household food distribution

The total energy intake provided by the meals served at home was 1,902-2,048 kcal for men, 1,854-1,755 kcal for women, and 887-1,036 kcal for children 2-5 years old (table 5). Comparison of the fulfilment of energy requirements did not show a significant difference between the two seasons for any group; however, there seemed to be a trend for the adequacy of fulfilment to decrease for the adults during the pre-harvest season but to increase for the children. The level of fulfilment for the children was significantly higher than for the men in August.

No differences were found in the contributions of the different food groups to total energy intake among men, women, and children (table 6). Cereals supplied the largest part (about 80% in April, about 67% in August). Pulses came next, contributing 9% in April and about 14% for the adults and 20% for the children in August. Tubers did not contribute in April, but contributed about 6% to the energy intake of the adults in August. The children got only 2% of their energy from tubers. The contribution of fats and oils was low: about 4% in April and about 7% in August. Meat or fish, vegetables and fruits, nuts, and condiments did not contribute more than 1%-3% per group.

TABLE 4. Relative contribution (percentages) of food from different sources to the total energy and protein intakes shown in table 2

Source Energy Protein
Aug Dec Apr Aug Aug Dec Apr Aug
'90 '90 '91 '91 '90 '90 '91 '91
Own stock 61 63 65 53 68 65 70 55b,c
Own field 6 11 2a,b 5 7 12 1a,b 7c
Wild resources 9 7a 6a 8b 2 2 2a 1c
Gifts 15 13 13 20 14 12 12 21b,c
Purchase 7 5 11 12 9 8 12 14
Other sources 2 2 3 2 1 2 2 2

a. Different from August 1990, p c .05.
b. Different from December 1990, p < .05.
c. Different from April 1991, p < .05.

TABLE 5. Mean daily energy and protein intake from meals served at home and adequacy of fulfilment of requirements for men, women, and 2-5-year-old children in 25 households in two seasons-Manta, Benin.

  N Energy Protein
Intake (kcal) Adequacy (%) Intake (g) Adequacy (%)
Mean SD Mean SD
Intermediate season, April 1991
Men 25 1,902 667 66 57 23 105
Women 25 1,854 611 74 55 20 92
Children 41 887 311 71 27 11 163a,b
Pre-harvest season, August 1991
Men 25 2,048 871 61 60 27 113
Women 25 1,755 586 67 51 18 93
Children 31 1,036c 300 79d 32e 12 186a,b

a. Different from men, p < .001.
b. Different from women, p < .001.
c. Different from April, p < .01.
d. Different from men, p < .05.
e. Different from April, p < .05.

TABLE 6. Relative contribution (percentages) of different food groups to the energy intakes shown in table 5

  April August
Men Women Children Men Women Children
Cereals 80 81 78 67a 68a 66a
Tubers 0 0 1 6 6 2
Meat/fish 2 2 2 2 2 1
Pulses 9 9 9 13 14 20a
Vegetables/fruits 1 1 1 1 1 2
Fat 4 4 5 7b 7a 7
Nuts 1 2 2 1 0b 1b
Condiments 2 2 3 2 2 2a

a. Different from April 1991, p < .05.
b. Different from April 1991, p < .01.

Protein intake in April and August ranged from 57 to 60 g for men, 55 to 51 g for women, and 27 to 32 g for children (table 5). Requirement fulfilments exceeded 100% for men and children in both seasons; only women received their requirements, by 7%-8%. Children received significantly more in proportion to their requirement than either men or women in both seasons. Protein intake came mostly from cereals (68%-71% in April, 59%-62% in August) and then pulses (about 16% in April, and about 25% for adults and 32% for children in August) (table 7). The contribution of meat and fish was low (7% in April, 3%-5% in August). Vegetables, fruits, nuts, and condiments contributed little to protein intake (2% each).

TABLE 7. Relative contribution (percentages) of different food groups to the protein intakes shown in table 5

  April 1991 August 1991
Men Women Children Men Women Children
Cereals 71 71 68 62 61a 59
Tubers 0 0 1 5 5 1
Meat/fish 7 6 7 5 3a 3a
Pulses 16 16 17 2 3 3
Vegetables/fruits 2 2 2b 2 3 3
Nuts 2 2 2 1 1a 1a
Condiments 2 3 3 2 2 2a

a. Different from April 1991, p < .05.
b. Different from men, p < .05.

Contents - Previous - Next