Contents - Previous - Next

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

Nutritional effects of export-crop production in Papua New Guinea: A review of the evidence

Peter F. Heywood and Robin L. Hide


Survey of effects by crop and location

Copra and cacao

Coconuts (Cocos nucifera) were an important source of food and drink in the traditional lowland subsistence system. The German administration introduced plantations for the production of copra in East New Britain Province in the 1880s. The dominant position of the plantation sector was maintained through the 1960s, when estates still accounted for approximately 70% of total copra production. Since that time the relative importance of the plantation sector has fallen to the point that small-holders now account for 58 % of production (table 3). There is marked geographical concentration of production, with East New Britain, North Solomons, New Ireland, and Madang provinces accounting for 77% of total production in 1983 (table 4).

TABLE 3. Small-holder and estate production of principal export crops, 1983




No. of small-holders

% of production

Trend 1979-1983a

No. of employees

% of production

Trend 1979-1983a




+ 4







+ 2







+ 1




Oil palm



+ 11













1 ,300



Source: Ref. 39.
a. Average percentage increase or decrease per year in production. b. Negligible.

Although cacao (Theabroma cacao) was introduced by the German administration in the early 1900s, production remained at very low levels until the rapid expansion of the 1950s stimulated by high prices and the vigorous extension efforts of the Australian administration. In the early 1970s, estates accounted for 75% of total production. The rapid expansion of small-holder production in the 1970s meant that by 1983 it was responsible for 69%. In 1983 two provinces, North Solomons and East New Britain, accounted for over 80% of total production.

TABLE 4. Estimated small-holder and estate incomes for principal export crops by province, 1983 (K '000)

Region and province




Oil palm


Income per capita (kina)











New Guinea islands
North Solomons








East New Britain








West New Britain












New Ireland















Western Highlands






Eastern Highlands
















Southern Highlands






New Guinea mainland










East Sepik


















West Sepiks




























Milne Bay


































Source: Ref. 39.

The spatial imbalances between provinces in the production of copra and cacao are strongly reflected in the pattern of average per capita incomes derived from these crops. Agricultural incomes in North Solomons and East New Britain are considerably higher than elsewhere, either in the New Guinea island region or in mainland locations where these two crops are important. Since provincial data on expenditure patterns are lacking, consideration of possible nutritional effects of this level of cash cropping requires examination of village case studies.

North Solomons

Two studies from opposite sides of Bougainville Island, in North Solomons Province, show significant differences between mainly copra- and mainly cacao-producing areas. In Nagovisi, in the southwest, intensive cacao planting began as late as 1960, and thus significant crop sales date only from the mid-1960s [12]. By 1973 the average household had a cash-crop holding of 1.43 ha of only cacao, 0.36 ha of only coconuts, 0.2 ha of mixed cacao and coconuts, and 0.05 ha of coffee. This compares with an average of 0.2 ha of food gardens cultivated annually.

From cacao sales alone, Nagovisi cash incomes averaged a modest $10 per household per month during 1968-1971. They varied widely, however, from less than $1 to nearly $30. Of an average retail expenditure of $4.33 per household per month, $3.72 (86%) was spent on food. Dependence on imported purchased foods was therefore relatively low; indeed, it was estimated that, on average, imported foods provided the basis of only 8% of meals.

The cash-crop situation of Teop, on the north-east coast of Bougainville, contrasts strongly with that of Nagovisi. Copra was sold from the early years of the century and retained its dominant position into the 1970s, when the average household cash-crop holding included some 600 coconut and 400 cacao trees, in all approximately 5.6 ha [42]. Incomes in 19731974 were considerably higher than in Nagovisi.

Average expenditure on imported food was a substantial $30 per household per month. As in Nagovisi, there was considerable variation among households, from $8 to $66. With this level of expenditure, purchased food accounted for 27% of the energy consumption and 23% of the protein consumption of the average adult.

The general picture of the size of cash-crop holdings and levels of expenditure provided by these village-level studies is confirmed by broader investigations [43-45]. Unfortunately, despite the length and intensity of the colonial period in the island region, little information is available relevant to assessing the effect on the nutrition of children. A provincial nutrition survey in 1980 sought associations between nutrition status and socio-economic factors such as cash-crop income, but no obvious relationships were found in the preliminary analysis (G. Marks and C. Kanawi, personal communication, 1981). However, subsequent analysis found some general relationships [27]. Several studies of adults, both cross-sectional and longitudinal, allow some assessment of the general effect of involvement in the cash economy, and the associated provision of government services, on nutrition.

Two longitudinal studies of secular trends in anthropometric variables in North Solomons are particularly relevant. One [46] measured heights and weights of adult men in the Siwai area in 1939, and the other [47] measured adults in the same area a quarter of a century later, some of whom had been measured earlier [46]. Comparison of the two cross-sectional surveys and of the measurements on the same individuals 28 years apart indicated significant secular trends in both height and weight.

A longitudinal study of adults and adolescents was conducted over the periods 1966-1970 and 19781980 in three areas of North Solomons with different levels of involvement in the cash economy: Nasioi, with the longest and most intense involvement; Nagovisi, where the involvement was intermediate (see above); and Aita, where the degree of involvement at the time of the study was minimal [48, 49]. For stature, the results in Nasioi indicate a longstanding and continuing substantial increase in stature in both sexes, but particularly in women. In Nagovisi the trend was significant in the younger ages only, particularly in females. No secular trend in height was evident in Aita.

In both Nasioi and Nagovisi there was evidence of weight gain in younger males and into middle age for women. In Aita no evidence of weight gain was seen in males, and females showed the downward trend in weight with age typical of traditional societies in Papua New Guinea. Skinfold measurements showed that these weight patterns were associated with increasing fatness with age in both sexes in Nasioi and Nagovisi, and in Aita a decrease in fatness with age in both sexes. The adolescent data (10-20 years) were not analysed by region. Average adolescent male stature increased slightly (0.5 cm, not statistically significant) over the 1970s, whereas that of females increased by 1.3 cm.

East New Britain

On the Gazelle Peninsula of East New Britain the high level of participation in cash cropping is complicated by significant cash income received from wage employment, particularly in the provincial capital of Rabaul. This is not recent, as shown by three studies of separate periurban communities on the outskirts of Rabaul between 1955 and 1978. At Malaguna in 1955 [50] all households in a small sample included at least one wage-earner; most had some subsistence food gardens; and some also had cash-crop holdings. The average household was spending a high 11 14s monthly on food, but most energy was derived from traditional starchy staples rather than imported cereals (the weekly per capita consumption of the former was 6.05 kg, compared with 1.38 kg of the latter).

Five years later a survey of Matupit showed 40% of adult men in wage employment, significant cash cropping (about 200 coconut trees per grower, and first cacao sales the previous year), and a substantial proportion of food still home-produced [51]. Wage-earning households were spending 7 8s monthly on food, non-wage-earning households 6 0s.

A deepening trend of food dependence is suggested by a study of another periurban community, Pilapila, in 19771978 [52]. By this date nearly two-thirds of households had a wage-earner—indeed, 15% had two; just over one-fourth of monetary income was derived from cash cropping; and, most significantly, 40% of all households were without gardens and were thus purchasing most of their food. An expenditure survey showed food constituting 55%-60% of household expenditure, with absolute outlays of between K 40 and K 84 monthly for food purchases.

Other accounts of rural Gazelle communities in the early 1960s located farther from Rabaul showed smaller proportions of income derived from wages, very substantial cash-crop holdings, in which coconuts were dominant, and estimates of 25%-50% of cash income used for food purchases [53, 54].

Although longitudinal anthropometric data for Gazelle are unfortunately lacking, the region was included in a cross-sectional study as part of a general health survey in the highland, mainland, and island regions [55]. Within the island region, adults living in villages classified as highly urbanized (indicating more involvement in the cash economy, a more Western diet, and possibly different levels of activity) were heavier and had greater skin-fold thickness and higher serum cholesterol levels than those living in more traditional villages. However, a recent longitudinal study of the coastal fishing community of Pere, on the island of Manus (parts of Manus, but not Pere, were included in the earlier survey), found no evidence of secular change in adult stature between 1954 and 1980 or in the stature of children between the mid1960s and 1980 [56].

In the Kilenge area at the western end of the is land of New Britain, a major expansion of coconut planting was associated with declines in food security, nutritional intake, and dietary diversity, as well as environmental degradation and economic dependence [57, 58]. Unfortunately, the case for negative nutritional effects depends only on qualitative assessment and cannot be evaluated on present evidence.

Other lowland areas

Evidence from mainland copra- and cacao-producing areas is scanty. Although provincial-level production and income data indicate lower levels of cash-crop involvement, there are clearly some local exceptions. Karkar Island, just off the north coast of Madang, appears to be one. In the mid-1960s official survey data gave mixed coconut and cacao holdings of 1.3 ha per grower, suggesting an average household figure of just under 1 ha [11]. A dietary study in one village in 1969, made as part of a wide-ranging series of investigations under the International Biological Program, showed store-purchased foods accounting for 21% of mean daily adult energy intakes and 29% of daily protein intake [59]; the figures for children were similar [60].

As part of the same study, the relationship between socio-economic status (SES), assessed by economic activity, occupation, education, and marital status, and a range of medical variables was examined with a cross-sectional sample of Karkar adults [61]. No differences were recorded in height among groups of high, medium, or low SES. Weight, weight-to-height ratio, and alcohol use were all greater in the high-SES group. In contrast, those of lower SES had higher rates of respiratory and skin diseases, anaemia, and malarial parasitaemia.

A restudy of Karkar in 1983-1984 showed that adults of both sexes had on average become slightly (and only just significantly) fatter, and had a greater central deposition of fat, than those measured in 1969 [62]. Increases in adult weights and heights were also reported between 1969 and 1984-1985 [63]. Among the Amele on the mainland opposite Karkar, birth weights increased from approximately 2.6 to 2.9 kg between 1957 and 1977 [64].

There are also suggestive data from the province with the lowest per capita export-crop income, Western Province. In a cross-sectional study of diets and anthropometry in several Gidra-speaking villages in the Oriomo-Bituri area, on the mainland opposite Daru, adults in the coastal village of Dorogori were 2-3 cm taller and 511 kg heavier than the adults of the less accessible inland village of Wonie [65, 66]. Although data on cash cropping are not given, the diets of the two villages contrast markedly. At Dorogori 45% of the average daily energy intake (mainly rice) and 39% of the protein intake (mainly tinned fish and meat) were purchased, whereas in Wonie the comparative figures were only 6% and 9% [6669].

At inland Kaiapit, in Morobe, comparison between two sets of weight-for-age data on children under three years old collected in 1958-1960 and 1965-1967 showed increases of between 0.2 and 0.5 kg over the whole age range [70, 71]. Elsewhere, evidence from Gulf Province showed a trend for birth weights to increase over the past two or three decades [72] but no change over the period 19741980 [73]. However, a small birth weight series from Kwato Island, in Milne Bay, reported no trend of increase over the period 1936-1978 [74]. At Angugunak, in West Sepik Province, where there is little cash cropping [75], a comparison of weights of children under five years of age in 1967 and 1976 showed no significant change [76]. There was also no change in children's weight-for-age status between 1976 and 1983 [77].

Marked variation exists throughout the island and mainland coastal regions in the extent of small-holder cash cropping of coconuts and cacao. Patterns of cash expenditure similarly show a wide range of dependence on purchased foods. Before 1975, figures as high as 20%-30% of energy and protein intakes from retail-purchased foods are reported for some rural areas, with probably higher proportions from periurban communities, where cash incomes tend to be greater due to wage employment. In the last decade this trend appears to have continued, with higher figures appearing in parts of both East New Britain and Western provinces. Both coconuts and cacao are relatively land-extensive, and the average size of household holdings in favourable locations is several hectares. Where population density is high, competition with food crops for land is therefore a real possibility (i.e., Gazelle, parts of North Solomons, and small offshore islands such as Karkar and Bali, in Madang and West New Britain respectively). Some anthropometric evidence indicates trends for increases in weight, height, and birth weight in recent decades, particularly for areas with significant cash-cropping (North Solomons and Madang). Locations where no change is evident are those with little or no cash cropping, such as parts of Gulf, Milne Bay, and West Sepik.




The most important export crop in Papua New Guinea is coffee. As with cacao, it was first introduced by the German administration late in the nineteenth century, but the major expansion in planting in the highland region did not occur until the 1950s. Arabica coffee, grown throughout the highland provinces and the elevated parts of Morobe Province, constitutes 90% of the crop. Robusta coffee dominates the lowland coastal regions but accounts for less than 10% of production. This split between the two varieties also coincides with geographical concentration: over 80% of coffee comes from three of the five highland provinces—Western Highlands, Eastern Highlands, and Simbu. Seventy-five per cent of production is by small-holders, and the proportion is increasing. Whereas before 1975 the division between plantation and small-holder modes of production mirrored the colonial ethnic divide between foreign and national ownership, this is no longer generally the case. In Western Highlands the "countryside is being transformed into areas of capital-intensive high development cash cropping run by business management concerns on behalf of a new elite of national owners, surrounded by areas of less intensively farmed small-holdings occupied by the bulk of the people" [78]. In both Western and Eastern Highlands recent government policy has also supported the development of 20-ha coffee holdings or mini-estates that share more of the characteristics of plantations than normal small-holdings [79]. Unfortunately, the income and nutritional consequences of these major developments of the last ten years have not been monitored, and most evidence for understanding the nutritional effects of coffee production predates the 1980s.

Cash incomes from coffee in the major highland producing provinces did not reach significant levels until the mid-1960s. This meant, given the dominance of the crop as the main source of cash for most highlanders, that purchases of imported foods remained modest into the early 1970s [80-82]. By this time the general pattern of small-holder production was apparent: household holdings averaging 0.1-0.3 ha [83], most production occurring during a regular annual flush between approximately May and September, and consequently a pronounced seasonality to regional patterns of expenditure.

Data for Simbu Province, where population densities are among the highest in the country, have been analysed on several occasions for evidence of competition between coffee and food crops leading to a decline in food production [11, 13, 84-86]. All concluded that the current small holdings of coffee per capita represented no immediate threat to continued food production.

A major change in income levels and expenditure patterns occurred throughout the highland region during the coffee boom of 1976-1977: by mid-1977 coffee prices were six times higher than their 19711972 levels [83, 87]. Data from three village-level studies show the extent of the changes in incomes and expenditure.

In 1977 the average household coffee holding in the Eastern Highlands village of Kapanara was estimated at 0.18 ha [22], which provided a monthly cash income of K 18.4 per month, or 71 % of the estimated total cash income. Food expenditure was not recorded, but an estimated 35% of income, or K 9 per household per month, was spent on alcohol. Farther west, in the Daulo region of Eastern Highlands, where the average household held 560 coffee trees, or approximately 0.2 ha [88], average monthly income in 1977 was a high K 73.8 per household, of which K 31.4 (56%) was obtained for coffee [89]. Monthly expenditure averaged K 75, with food accounting for 36% (K 26.5), beer 20%, gambling 18%, ceremonial payments also 18%, and tobacco and betel an additional 9%.

Even higher levels of expenditure were reported in a 1978 study of the Simbu village of Yobakogl: K 128.7 a month per household, of which food accounted for K 30.7 (24%), alcohol 11%, gambling 16%, gifts 18%, tobacco and betel 6%, transport 13%, and other items a further 11% [19]. Households in Yobakogl could be divided into two groups: those for whom coffee income accounted for less than 50% of total income, with an average coffee holding of 0.14 ha, and those for whom coffee provided the major part of their total income, with smaller average holdings of 0.11 ha. In 1981, when food consumption was studied intensively in this village (but coffee prices had fallen considerably), retail purchases of food accounted for 26% of the energy and 35% of the protein intakes of adults [21, 90].

Several anthropometric studies of both adults and children in the highland coffee area are relevant to the question of the nutritional effects of cash cropping. The best documented sequence comes from the northern part of Simbu. In the early 1960s the growth of village children from various Simbu locations was compared with the growth of children living in the provincial capital [91, 92]. The parent(s) of the latter children were in wage employment and received rations including rice, sweet potatoes, wheatmeal, and tinned meat and fish. In both studies the children living in the capital were longer and heavier than those living in the villages, with some indication that birth weights in the villages were lower. Similarly, comparisons of adult weights in urban and rural communities found the former to be heavier [93].

In 1956 growth studies were performed on village children in Simbu at a time when there was virtually no cash income, and no cash expenditure on purchased foods was reported [94]. Two subsequent studies [95; 21, 90] were carried out with the explicit aim of comparing the results. By 1975 children under five years of age were approximately 1.5 kg heavier at all ages than in 1956 [95], and in 1981 for children above six months these differences had been maintained or increased slightly [21]. By the latter date, as described above, increased cash incomes had resulted in considerable expenditure on foods.

Significant increases in the mean weights of adult men and women were seen in 1975 [95]. The later study confirmed this trend for women and indicated that mean weight had increased further [90]. The mean difference between 1956 and 1980 was more than 5 kg. In addition, the average birth weight in north Simbu increased by about 200 g between the mid-1950s and the early 1970s [90]. During the 1970s, however, there was no subsequent change. Eight years later, in 1988, birth weights in the north of Simbu had remained stable during the first half of the 1980s, but the stature and weight of adult women had continued to increase in the one district examined (Gumine) [96].

Information on changes over time in other highland locations is growing. In Enga Province, significant increases in weight (approximately 1 kg overall) of children under five years old were seen between 1958 and 1977 [97]. At Lufa, in Eastern Highlands, adult weights and heights increased between 1967 and 1984 [63]. In Bundi, on the northern highland fringe in Madang Province, significant increases were found in both the weight and length of children under ten years old between 1967 and 1984 [98]. In addition, children under five years old of Bundi parents who were raised in town were significantly heavier and longer than their age mates living in the rural Bundi homeland. More details on change in adolescent growth showed that rural Bundi children 5-20 years old had significantly larger body sizes in 1984 than in 1967 [99]. The tempo of growth in height and weight increased as well, except for height in men. Urban-rural contrasts showed urban males of Bundi parents were taller, heavier, and broader than their rural counterparts 1115 and 16-20 years old. For females, the most significant difference in size was for those 11-15 years old. These findings are related to improved nutrition, which increased childhood growth and the rate of maturation.

Comparison of a regional anthropometric survey of the Nembi Plateau in Southern Highlands in 1980 with an earlier survey in 1978 showed minor improvement in nutrition status in the central and northern areas of the plateau (with more economic opportunity and better access), but not in the south, where the settlement was below 1,600 m and nutrition status poorer [100]. A subsequent anthropometric survey of the whole of Southern Highlands in 1983 [101] showed a positive relationship between the nutrition status (weight for age and weight for length, but not length for age) of children less than five years old and an index of cash-earning activity (assessed by village; small-holder coffee was one of the most important cash sources).

It has been suggested that the nutrition of small-holder coffee growers (and/or their food security) deteriorated as a result of the land and labour requirements of coffee production competing with the needs of food production [16, 19, 22, 95, 102]. (While Lambert's analysis [16] was disputed by Hide [20; see also 18] and his assessment of Sinasina nutritional effects replaced by that of Harvey and Heywood [21, 90], nearly a decade later Longhurst [4] was still referring to the summary discussion [16] as the main source of a study showing the negative nutritional effect of coffee.) Critical examination of this hypothesis [103] concluded that then-current highlands evidence, which included data on sweet potato planting rates, sweet potato price series, and the occurrences of food shortages, did not show any general associations between coffee production and rural food production.

Analysis of data from the 1982-1983 national nutrition survey from part of the highlands, and to a lesser extent the highland fringe, shows positive associations between indexes of household cash cropping and improved child growth (C. Jenkins and B. Zemel, personal communication, 1990). Using data from the three westernmost provinces in the highlands (Enga and Southern and Western Highlands), significant relationships were found between the growth of both male and female children (weight and height) and household participation in coffee growing and marketing food crops. In the case of boys, growing cardamom and participating in other enterprises such as selling betel nut and artifacts were also significantly associated with growth. In separate analyses of data from the mid-altitude zone of the northern fringe highlands and the lowland Sepik plains, similar relationships were seen in the former (although the relationships for girls were weaker) and none in the latter.

Dietary information from the Eastern Highlands village of Kapanara for 1977 was used to address questions relevant to the continued nutritional effects of agricultural change: whether increased cash income after commodity production is well established will continue to result in dietary improvement, and what factors affect the consumption of subsistence foods [104]. For a small sample of households the frequency of consumption of purchased foods (largely tinned fish and rice) was not related to household cash income. There was, however, a strong positive relationship between beer purchases and income. The frequency of subsistence food consumption was related to indexes of total garden area and the range of environmental zones occupied by gardens. Thus issues related to the nutritional effects of cash cropping and dietary change are best regarded as open and subject to changing circumstances.

The general picture in the highlands is of an initial 15 years, between about 1955 and 1969, during which coffee growing spread widely, cash incomes grew from nothing to modest levels, and dietary change began slowly and increased to low but regular consumption of imported foods [105]. Significantly, in the case of Simbu, increases in birth weights, the growth of children, and the weights and heights of adults were already apparent either by the end of this period or by 1975. An initial increase in the growth of Enga children also took place in this time. The following period, from 1970 to the present, was marked by a substantial rise in incomes associated with the coffee boom, and a major increase in the levels of consumption of purchased foods. Growth increases, possibly beginning earlier, occurred in Eastern Highlands, Bundi, and Southern Highlands. In Simbu, although this period has not seen continued increase in birth weights (but the current average weights are among the highest in the country), secular trends in adult women's stature and weight have been maintained in at least one district.


A 1983-1984 study of a village in Maprik District, East Sepik Province, found a positive relationship between the size of producing robusta coffee holdings and the nutrition status of children in a sample of 23 households [106, 107]. Cash crops (cacao had recently been added to coffee plantings) at this time occupied some 10% of total village land, but there was no evidence that food gardens had been displaced to distant or inferior sites.


Oil palm

The development of the oil palm industry began at Hoskins, in West New Britain, in 1967 and from the beginning was on a nucleus estate basis. Subsequent developments at Bialla, also in West New Britain, in 1972 and at Popondetta, in Oro Province, in 1976 used the same basic structure, as did new projects currently under way in Milne Bay and New Ireland. The government and a private company each hold 50% of the shares, with the company also providing management and marketing services. In 1983 the total area of the nucleus estates was 12,100 ha and of small-holders 16,600 ha. The total number of small-holders at the 1980 census was 4,500, and an additional 5,000 labourers were employed by the estates.

When the Hoskins small-holdings were surveyed in 1975 [108], the blocks had either 3.2 or 4.0 ha under oil palm, and an additional 2.8 or 2.0 ha available for food production, of which an average of 0.42 ha was currently cultivated. Home production of food was therefore substantial. Between 1976 and 1981 average household oil palm incomes ranged from K 205 a month in 1978-1979 to K 59 in 1980-1981. A small survey in 1982 [109] reported very low oil palm incomes (K 36 per household per month), with food sales at local markets adding an extra K 20 a month. Thus in 1982 market sales of garden produce averaged some 40% of household cash income. In the same year, retail-store expenditure on foods was an estimated K 17 per household per month, with an additional K 5 a month spent on market purchases. Rice and tinned fish, as elsewhere, dominated the list of purchased foods. Most food purchases were made from market income (i.e., under the control of women), not from oil palm income, which was mainly controlled by men [109].

Although at least three anthropometric surveys of women and children were made at Hoskins [110112], differences in sample size and selection and lack of an adequate comparative population prevent firm conclusions about trends in nutrition status on present evidence. To summarize, in 1975, 42% of 75 settlers' children under five years of age from one settlement subdivision (Buvussi) and 43% of 70 oil palm labourers' children were under 80% of the standard median weight for age [110]. Later this settler percentage was contrasted with 23% for village children from the same area [16, 113], but this comparison was not made in the original report, and the origin of the village figure and the sample size are unknown. Elsewhere it was reported that the results of the 1978 national nutrition survey (NNS), which was based on maternal and child health clinic data, showed 51% of 974 children in the oil palm settlements were malnourished, compared with only 15% of the 942 children seen in village clinics outside the settlement area [114].

Checking this substantial difference, a late 1978 survey of 136 settler children under five years of age from five subdivisions (but not, unfortunately, Buvussi) found 37% under 80% weight for age, slightly lower than the 42% in the 1975 survey but substantially lower than the 51 % reported by the NNS [111]. But six and even twelve years later it was Lambert's undocumented figures rather than Singleton's that were still referred to as evidence of the negative nutritional effects of oil palm [115, 116]. In 1982 only 17% of 136 oil palm labourers' children were under 80% weight for age [109, 112]. Although this was a considerably lower proportion than that reported seven years earlier, the comparison was not made. The numbers of adult women seen at these surveys are too few and lack sufficient control of ethnic origin to allow meaningful comparison. One baseline set of anthropometric data only is available for the Higaturu scheme at Popondetta [117, 118]. Socioeconomic information from the scheme, however, is as yet minimal [119, 120].

Given the major role played by government agricultural investment in these oil palm projects, the lack of reliable information concerning their nutritional effects on children and adults is unfortunate. Uncritical acceptance of inadequately documented negative nutritional effects is apparent [121, 122].


Other crops


Rubber was the major cash crop in the Gavien resettlement project in East Sepik Province, which for over a decade has received wide publicity as an example of the negative effects of cash cropping on nutrition [2, 3, 16, 113, 123125]. After substantial refunding and upgrading between 1977 and 1985, the project was investigated in 1986. The nutrition status and diets of settler children and mothers belonging to three ethnic groups living on the project (Grass Country, Wosera, Middle Sepik) were compared with those of residents in the three home areas. It was found that the nutrition status of the children from the Grass Country and Wosera was generally higher on the project than that of their home area counterparts; the nutrition status of the Middle Sepik children was similar in both localities [35]. Maternal nutrition status differences between the project and home areas were restricted to Wosera women, who had higher skin-folds and arm circumferences on the project. Diets also generally were improved on the project. Thus resettlement, although focused on cash cropping, need not necessarily result in decreased food diversity and nutrient intake [36].

Using only the sample resident on the project, the effects of such household-level variables as cash-crop income, expenditure on food, and the number of food crops planted in gardens on various indexes of both childrens' and mothers' nutrition status and dietary intake were examined. The garden-food index was not related to children's nutrition status, and expenditure on foods was only positively related to arm circumference. However, cash-crop income was positively related to the nutrition status of children. In the case of mothers, nutrition status (BMI) was not related to cash-crop income but was positively correlated with the garden-crop index. Although this study was carefully controlled for ethnic group and season, the samples were small and the data on garden crops, cash-crop income, and household expenditure are of variable quality. Thus the results should be interpreted with caution.


A major study undertaken in 1987-1988 at Karimui, at 1,200 m in the south of Simbu, with the collaboration of the International Food Policy Research Institute, focused on the effects of wage employment in cardamom cultivation on household cash and subsistence incomes, the effects of the level and source of income on food consumption patterns, and the effects of these changes on the growth of children under five years old [38]. The socioeconomic environment was unique for Papua New Guinea in the rapidity and scale of cardamom development and the scale of women's employment in the enterprise. The study involved 90 households in three settlements, with a total of approximately 150 children under five years old. The weight and length growth velocities of these children in two six-month periods were used as indexes of nutrition status and related to a number of independent variables.

The results were complex, as the study was carried out at a time when the cardamom enterprise began to fail, the level of wage labour fell, and cash income and expenditure declined. During the first six months (August-December 1987) average cash incomes fell threefold [38], whereas during the second (January-July 1988) incomes were stable but the proportion derived from wages rose and, while total expenditure remained stable, the proportion of expenditure on food rose. In terms of energy availability, the first period was apparently worse than the second: during the second, food was more available, the proportion obtained from garden staples fell from 80% to 69%, the proportion purchased fell from 14% to 11%, and the proportion from other sources rose from 6% to 19%.

Regression analyses showed that greater involvement of households in the cash economy was associated with higher levels of total income, apparently because the households were able to maintain a subsistence income while engaging in wage labour. Although greater cash participation diversified the foods consumed, it had little impact on total energy intake. The budget share allocated to buying food declined with increasing cash income, but the consumption of purchased foods rose.

The nutritional effects on the growth of children under five years old were complex. Leaving aside factors such as age, sex (no difference), and a child's weaning and initial growth status, growth was negatively (but not significantly) affected by the frequency of two broad categories of illness, fever and diarrhoea. Growth was not related to maternal care as indexed by such measures as the frequency with which a mother cared for her child, worked for wages, or worked in food gardens. It was not significantly affected by total household income, the proportion of income from wages, or the proportion from subsistence production, although the trend was for an overall negative effect of both income and the wage proportion of income. Growth was also not significantly affected by total expenditure, the proportion of cash expenditure on food, or the proportion of food derived from own production. Again, there was an apparent trend for overall negative effects of increasing cash expenditure and increasing proportion of food expenditure, but the effects were less in the second period.

Child growth was not related (either by weaning status or period) to the total availability of food energy at the household level. When food was disaggregated by source, the percentage of total energy from purchased foods had a positive overall effect on both weight and length growth, especially weight. The effect on weight reversed between the two periods: it was negative in the first and positive in the second. The effect was also greater for children in the process of weaning than for pre-or post-weaning children. The proportion of food from gardens had similar but less significant effects. Further disaggregation of food by type showed that the proportion of total energy available derived from staple crops (root crops, banana, sago) had an overall negative effect on child growth, particularly weight. Again, however, the relationship changed with period: from positive in the first to negative in the second. The effect also varied with weaning status, being negative for weaning children and positive for pre- or post-weaning children.

For rubber, although the results are not definitive, there are clear indications that children living on one small-holder settlement scheme had better nutrition status than those living in the home areas from which the settlers were originally drawn. In the case of cardamom, because of the financial difficulties of the project and the short time frame, no clear nutritional effects of wage employment on one estate were identified.

Contents - Previous - Next