the commercialization of agriculture and household-level food security: the case of southwestern...

7
World Development, Vol. 16, No. 9, pp. 10751081,1988. 0305-75OW88 $3.00 + 0.00 Printed in Great Britain. 0 1988 Pergamon Press plc The Commercialization of Agriculture and Household-level Food Security: The Case of Southwestern Kenya EILEEN KENNEDY and BRUCE COGILL International Food Policy Research Institute, Washington, DC Summary. - Results from a 1984/85 study in South Nyanza, Kenya to evaluate the effects of shifting from maize to sugarcane production on household-level food security are presented. Household food security is not jeopardized by the entry into commercial agriculture. The area allocated to subsistence food production is similar for sugar- and non-sugar-producing house- holds. Incomes of sugarcane producers are significantly higher than non-cane producers and a portion of this incremental income is used to increase household caloric intake. The conventional “wisdom” that the transition to commercial agriculture causes a deterioration in household-level food security is not borne out by the results from the present study. 1. INTRODUCTION The issue of the appropriate role of export crops/cash crops’ in many developing countries is politically volatile. On the one hand, many governments are encouraging the increased pro- duction of cash crops as a means of generating and saving foreign exchange earnings while, on the other hand, stressing the production of food crops for domestic consumption. At times, these goals appear to be in conflict. In Kenya, there has been some concern that in areas with increased cash cropping, particularly increased sugarcane production, household food security has been jeopardized. This concern is re- flected in a 1981 Kenyan National Food Policy Paper which indicated that particular attention be given to safeguarding the diet of the small farmer switching from maize to sugarcane production.’ Relatively little is known, however, about the consumption and nutritional effects of the process of commercialization of agriculture. This is somewhat surprising given the importance of cash crops/export crops in the economies of many developing countries. A recent review of the income and nutritional effects of cash crop production suggests mixed results;3 while some studies show a negative effect of cash crop pro- duction on consumption and nutritional status, other studies show a positive or neutral effect. A concern that cash cropping households have compromised the ability to meet their own food needs has placed food security as a major con- sideration in commercial agricultural projects. The present paper examines the effects of the commercialization of agriculture on household food security in southwestern Kenya. South Nyanza, Kenya was chosen as the study site for the research, in part because it has the highest birth to two-year mortality rate - 2161 1,000 - of any part of the country and is an area targeted for development assistance. South Nyanza is also an area where rapid commerciali- zation of agriculture has been taking place. Policymakers were concerned that the heavy emphasis on sugarcane production was causing a deterioration in household food security. 2. PROJECT DESCRIPTION AND SAMPLING In 1977, the newest of the sugar factories in Kenya was established - the South Nyanza Sugar Factory (Sony). The Sony factory obtained approximately 2,500 hectares of land from local 1075

Upload: eileen-kennedy

Post on 31-Aug-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

World Development, Vol. 16, No. 9, pp. 10751081,1988. 0305-75OW88 $3.00 + 0.00

Printed in Great Britain. 0 1988 Pergamon Press plc

The Commercialization of Agriculture and Household-level Food Security:

The Case of Southwestern Kenya

EILEEN KENNEDY

and

BRUCE COGILL International Food Policy Research Institute, Washington, DC

Summary. - Results from a 1984/85 study in South Nyanza, Kenya to evaluate the effects of shifting from maize to sugarcane production on household-level food security are presented. Household food security is not jeopardized by the entry into commercial agriculture. The area allocated to subsistence food production is similar for sugar- and non-sugar-producing house- holds. Incomes of sugarcane producers are significantly higher than non-cane producers and a portion of this incremental income is used to increase household caloric intake. The conventional “wisdom” that the transition to commercial agriculture causes a deterioration in household-level food security is not borne out by the results from the present study.

1. INTRODUCTION

The issue of the appropriate role of export crops/cash crops’ in many developing countries is politically volatile. On the one hand, many governments are encouraging the increased pro- duction of cash crops as a means of generating and saving foreign exchange earnings while, on the other hand, stressing the production of food crops for domestic consumption. At times, these goals appear to be in conflict.

In Kenya, there has been some concern that in areas with increased cash cropping, particularly increased sugarcane production, household food security has been jeopardized. This concern is re- flected in a 1981 Kenyan National Food Policy Paper which indicated that particular attention be given to safeguarding the diet of the small farmer switching from maize to sugarcane production.’ Relatively little is known, however, about the consumption and nutritional effects of the process of commercialization of agriculture. This is somewhat surprising given the importance of cash crops/export crops in the economies of many developing countries. A recent review of the income and nutritional effects of cash crop production suggests mixed results;3 while some studies show a negative effect of cash crop pro-

duction on consumption and nutritional status, other studies show a positive or neutral effect.

A concern that cash cropping households have compromised the ability to meet their own food needs has placed food security as a major con- sideration in commercial agricultural projects. The present paper examines the effects of the commercialization of agriculture on household food security in southwestern Kenya.

South Nyanza, Kenya was chosen as the study site for the research, in part because it has the highest birth to two-year mortality rate - 2161 1,000 - of any part of the country and is an area targeted for development assistance. South Nyanza is also an area where rapid commerciali- zation of agriculture has been taking place. Policymakers were concerned that the heavy emphasis on sugarcane production was causing a deterioration in household food security.

2. PROJECT DESCRIPTION AND SAMPLING

In 1977, the newest of the sugar factories in Kenya was established - the South Nyanza Sugar Factory (Sony). The Sony factory obtained approximately 2,500 hectares of land from local

1075

Page 2: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

1076 WORLD DEVELOPMENT

landowners to establish the factory and nucleus estate. The majority of sugarcane, however, is produced by smallholders or outgrowers under contract with the Sony Company. The out- growers’ program includes 6,000 farmers and approximately 6,000 hectares of land under con- tracted sugarcane.

The outgrowers’ program at Sony is similar to the outgrowers’ programs used in other sugar schemes in Kenya. Farmers are under contract with the factory; the factory agrees to purchase sugarcane from the outgrowers at the price that is prevailing at time of harvest. The price paid to the sugar producer is set by the government of Kenya.

The factory provides a number of services for a fee to the farmers involved in the outgrowers’ scheme. Typically, the factory will survey the farmland to identify the acreage most suitable for sugarcane production. In addition, the factory may also clear the brush, prepare the land, pro- vide seed and fertilizer, plant, plow, weed, cut the cane, and transport the final crop. The extent to which the factory provides these services varies depending on the individual needs of the farmer and the availability of inputs from the fac- tory. The overwhelming portion of crop inputs including labor are supplied by the factory.

The study sample includes a representative sample of sugar-producing households.’ The sugar company provided a master list of farmers ever in the outgrowers’ scheme. From the list, a random sample of sugar farmers weighted by sublocation was chosen.’ Each of the randomly selected sugarcane farmers had to have (1) at least one child five years of age or younger in the household, (2) less than 20 hectares of land, and (3) could not be a non-resident farmer.

Since the Sony factory is the newest sugarcane scheme in Kenya, it was still undergoing expan- sion at time of the study. This provided the opportunity to identify a cohort of farmers prior to entry into the outgrowers’ program and/or prior to the first sugarcane harvest and to col- lect baseline information on socio-demographic characteristics and health and nutritional status. This group of sugar farmers is called the “new en- trants.” Of the 181 sugar farmers in the study sample, 23% had not received payment for a sugar crop.

The remaining 77% of the sugar farmers had received at least one payment for the cane crop and are called the “sugar farmers.”

Once the sugar farmers’ sample was chosen, field staff identified the next nearest non-sugar farmers who met the same selection criteria.6 This approach ensured geographic similarity of sugar and non-sugar farmers. The nearest-

neighbor method of sampling non-sugar farmers essentially enables the use of the sugar farmers as a seed unit by mapping all neighbors who did not grow sugarcane. For each sugarcane contractor, mapping was performed on up to three non-sugar households, of which up to two were randomly selected.

Beginning in June 1984, data were collected from households every two months over an eight- month period on food production patterns and labor use, income and expenditure, food con- sumption, morbidity patterns, and nutritional status. It is important to emphasize that while 1984 was a drought year for Kenya, the project area was not as severely affected. It was the late arrival of the long rains, rather than an absolute deficit in rainfall, that created more of a problem for agricultural production. However, agricul- tural production for all households in the com- munity was lower than what would be expected in a normal rainfall year.

3. RESULTS AND DISCUSSION

Household food security can be thought of as the ability to provide adequate energy intake either from food produced directly by household members and/or through the availability of suffi- cient income to purchase food. The present paper examines the effects of sugarcane produc- tion on semi-subsistence production and total household income.

One of the longstanding criticisms of commer- cial agriculture has been that the amount of food available for home consumption is drastically de- creased when households became cash crop pro- ducers. Data in Table 1 would suggest this is not true. The area of land per adult equivalent used for staple food production is remarkably similar across the three different categories of agricul- tural households. Given this, it is not surprising that there are also no significant differences in either the number of days of calories in storage or the mean energy intake per adult equivalent in the household (Table 1).

In addition, average energy intakes for the new entrants, sugar farmers, and non-sugar farmers were analyzed within quartiles of farm size categories (Table 2). There are no significant dif- ferences across the three types of agricultural households in mean kilocalorie consumption in any of the farm size quartiles. What is more sur- prising is the lack of significant variation in energy intake as one moves from the smallest to largest farm size category. Even in the smallest farm size category, high priority is placed on attaining adequate food consumption.

Page 3: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

AGRICULTURE AND HOUSEHOLD-LEVEL FOOD SECURITY IN KENYA 1077

Table 1. Descriptive statistics for land allocation and energy intake for agricultural households for the long rains season,

May-August 1984

Type of household New Sugar Non-sugar

entrants farmers farmers

(N = 40) (N = 132) (N = 231) Area land in food crops/

adult equivalent (ha) 0.29* 0.27 0.28 Kilocalories in storaget

(number of days) 45.0* 57.0 44.0 Household energy intake

per adult equivalent* 2,822* 2,689 2,669

Source: IFPRI survey in South Nyanza, 1984/85. *No two groups significantly different. tAll-round average. Based on the number of days the food in storage would satisfy the energy requirements of the entire household (2,850 kcals-adult equivalent used as the standard). fBased on average for each household derived from one to four i4-hour dietary yecalls.

Table 2. Energy consumption per adult equivalent by farm size category

Farm size category (hectares/capita)

Energy intake/adult equivalent* New Sugar Non-sugar

entrants farmers farmers

0.0-0.082 hectare 2,478t 2,462t 2,519t 0.083-0.146 hectare 2,936 2,561 2,802 0.147-0.247 hectare 2,794 2,659 2,574 0.248 hectare or more 2,882 2,887 2,852

Source: IFPRI 1984/85 South Nyanza survey. *Mean energy intake based on all round average for each household. No groups significantly different. tNo significant difference between quartiles of farm size/capita.

Sugarcane and non-sugarcane producers do have different land utilization patterns (Table 3). Non-sugar farmers have a significantly smaller percentage of total farm area devoted to all crops when compared to sugar farmers. A major part of this difference is due to the growing of con- tracted sugar.

The larger percentage of area devoted to all crops by sugar farmers is reversed when just food or edible crop area is compared. During the long rains, sugar farmers have a significantly smaller percentage (36.0) of their land in food crops com- pared to non-sugar farmers,’ although, as already discussed, the area of land per adult equivalent for staple food production is similar for sugar- and non-sugar-growing households.

For almost all the study households, labor is more of a constraint to production than is land.8 This is one major reason why less than 60% of land, on average, for non-sugarcane producers is put into production.

The creation of the sugercane outgrowers’ scheme has changed the credit environment for the farmer; the availability of factory-supplied labor allows the sugarcane farmer to put a greater proportion of farm land into production without any additional demand for household labor. As shown in Table 3, the amount of house- hold labor supplied for all crops is similar for sugar farmers, new entrants, and non-sugar farmers despite marked differences in land area under cultivation.

However, even if subsistence production of staples decreases as a result of the shift into com- mercial agriculture, household food security need not deteriorate; if income gains as a result of cash crop production are substantial, a propor- tion of this incremental income could be used to counterbalance the decline in home production.

Table 4 presents a distribution of total income and income by source per capita for the study households. It is worth repeating that since 1984 was a year with lower rainfall than normal, the incomes of households in community may be lower due to decreased agricultural production. However, one would expect that agricultural households - sugar- and non-sugar-producing - would have been affected equally by the poor climatic conditions. While bearing in mind that absolute incomes may be lower than usual, the

Page 4: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

1078 WORLD DEVELOPMENT

Table 3. Land allocation and labor use patterns for agricultural households, long rains, 1984

Type of household New Sugar Non-sugar

entrants farmers farmers

Mean percent of all land under crops*

Mean percent of land under food crops§

Total household labor/l (person days/capita/ha)

Hired labor** (person days/capita/ha)

51.7t 66.9$,t 56.6$

36.47 36.0$ 52.l$,ll

12.4 8.3 13.8

1.7 1.9 1.2

Source: IFPRI 1984/85 South Nyanza survey. *Excludes fallows, pastures, and wooded areas. Multi-cropped plots counted only once. tT-test for sugar farmers and new entrants p < 0.05. ST-test comparison of sugar and non-sugar farmers p < 0.05. POnly edible crops included. Yield may at some time be sold for cash. IT-test for non-sugar farmers and new entrants p < 0.05. /IChild labor computed as 0.5 of an adult person day. **Does not include factory-provided labor.

Table 4. Income* and income by source for agricultural households, 1984

Activity

group

Agricultural income Used for own Non-agricultural Total consumption Marketed income income/ X Percent X Percent 8 Percent N capita

New entrants 728t 37 404 21 X24+ 42 42 1,956 Sugar farmers 748t 29 9422 36 901t 35 139 2,591§ Non-sugar farmers 822t 43 393 20 709t 37 231 1,924 Total sample mean 786 37 578 27 784 36 412 2,152

Source: IFPRI 1984/85 South Nyanza survey. *All values shown are in Kenyan shillings (Ksh), 16 Ksh = 1 US dollar. tNo two groups significantly different. *Sugar farmers have significantly (p < 0.050) higher marketed agricultural income per capita than other groups. §Sugar farmers significantly (p < 0.050) higher total income per capita than other groups.

main interest in this paper is the comparison of the relative incomes across groups.

Sugar farmers have significantly higher in- comes per capita than the non-sugar farmers. The new entrants to the sugar scheme who have yet to receive payment for a sugar crop have incomes/capita very similar to non-sugar farmers. Much of the differences in income between sugar and non-sugar producers is due to cane produc- tion.

Income from agricultural sales is significantly higher for sugar farmers than for either the non- sugar or the new entrant households. The agri- cultural sales income/capita for sugar farmers is 2 to 2.5 times higher than that of non-sugar far- mers. Much of this difference is due to the income from sugar production; 400 Kenyan shill- ings (Ksh) out of 942 Ksh, or approximately 42%) of marketed farm income is contributed by sugarcane.

Page 5: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

AGRICULTURE AND HOUSEHOLD-LEVEL FOOD SECURITY IN KENYA 1079

Alternatively, this can be presented in another way. Sugar farmers earn approximately 550 shillings/capita more from commercial agricul- tural income than non-sugar farmers; of this 5.50 shillings/capita, 73% is contributed by sugar. Sugar income makes a substantial contribution to household income.

Income from marketed surplus is similar for the new entrants and non-sugar farmers - non- sugar farmers have agricultural sales income/ capita of 20%; and for new entrants, 21% of income/capita comes from agricultural sales. In contrast, for sugar farmers, 36% of total income comes from agricultural sales.

The income/capita used for own consumption is remarkably similar across the three types of agricultural households; these results corrobo- rated the data already presented in Table 1.

The analyses of energy consumption thus far have been a static comparison across groups. This descriptive analysis does not allow us to identify the determinants of household caloric consumption. Table 5 presents results of a house- hold consumption function.’

As expected, the number of adult equivalents exhibits a positive, significant association with household energy intake (Table 5). Total house- hold annual income also shows a positive and sig- nificant effect on caloric intake. However, this income/calorie relationship, although significant,

Table 5. Regression of daily household kilocalories for total sample using income and calorie storage models

Variables

Income model* total

sample

Income 0 O?) Percent of income from sugar Percent of income by women

15:48 iii?;/+ -1.61

Adult equivalent 1508.17 (32.68) Maize price 365.83 HOH schooling 134.72 i:::;; Mother’s schooling 106.68 Round 2 dummy 53.27 i::;:; Round 3 dummy 957.89 Round 4 dummy 1,944.71 i:::;; R2 0.510 DF 1,190 F 125.9

Source: IFPRI 1984/85 South Nyanza survey. *Regression was respecified using income and income squared and total expenditures and total expenditures squared and results were similar. This model is based on agricultural households only. +T-statistics are in parentheses.

is weak. The marginal propensity to consume calories is very low. For the total sample, for each 100 shillings increase in annual income, household caloric intake increases by only two calories. From the data in Table 4, we saw that incomes in the sugar farm households that have received payment from contracted sugar sales were approximately 670 shillings per capita per year higher than other farming households. Using the model in Table 5, this incremental in- come would account for an increase in household caloric intake of approximately 134 calories per day or 18 calories per adult equivalent.

There is a round or season effect. Household caloric consumption is significantly higher in Rounds 3 and 4 which is the later post-drought, short-rains harvest period.

The schooling of the head of household has a positive effect on household energy intake. For each additional year of education, household consumption is increased by 134 calories. This education variable is probably acting partially as a proxy for household wealth.

In summary, household caloric consumption is positively and significantly influenced by total household income. Since participation in the sugar outgrowers’ scheme is associated with an increase in income, the energy intake of the household benefits, although the effect is small.

4. SUMMARY AND CONCLUSIONS

The effect of commercial agriculture on household-level food security has been a conten- tious issue. Critics of accelerating the production of export crops/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 influence on food consumption and nutritional status.”

The Kenyan government has been concerned about potential negative nutritional effects of cash crop production; national-level data from the late 1970s indicated that in some areas with increased sugercane production, preschooler nutritional status appeared to be deteriorating.” The South Nyanza study was initiated at the request of the government of Kenya to investi- gate the impact on income, production, and con- sumption of shifting from maize to sugarcane production.

The results from this study suggest some positive impacts of commercial agriculture on household income. The income of farmers in the sugarcane outgrowers scheme are 670 shillings per capita (approximately 25% of income) higher than the non-sugar farmers in the same region.

Page 6: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

1080 WORLD DEVELOPMENT

Much of this income differential is due to agricul- tural sales - in particular sugar; 73% of the difference in agricultural sales between sugar and non-sugar growers is attributed to sugar production.

The higher income associated with sugarcane production has resulted in a modest increase in household caloric intake. However, the income/ household calorie effect is small. The lack of a more robust effect of income on household caloric consumption may relate, in part, to who controls different types of income. In the study men and women not only have different sources of income but also different expenditure respon- sibilities. In this study, 79.0% of respondents indicated that men control the income from sugarcane production. In addition, 76.3% of re- spondents said that women are responsible for food expenditures.

For sugar farmers, incremental income re- ceived from sugar production is most likely to be spent on non-food items - in particular, housing and school fees.i2 While in the longer term these expenditure patterns may produce benefits, these may not necessarily produce food consumption benefits, particularly in the shorter term.

Two notes of caution. The cross-sectional design used in the present study is typical of what has been done in prior research and this approach is popular in the evaluation litera- ture. However, as with most cross-sectional approaches, there is no clear indication that the baseline consumption and nutritional status was similar for sugar and non-sugar producers. The sugar producers may have been wealthier, to begin with, and may therefore have done well regardless of whether or not they ultimately become sugar outgrowers.

It is encouraging that the new entrants group is similar to non-sugar producers with respect to income/capita because the new entrants re- present a group of farmers on whom we have

baseline socioeconomic and food consumption information prior to the sugarcane outgrowers’ program. A follow-up study covering the period December 1985 to March 1987 has been com- pleted. Data from this follow-up study indicate that the income of the new entrant group (3,931 Kenyan shillings per capita) is now significantly higher than the non-sugar cohort (2,70X Kenyan shillings per capita). Here again, much of the dif- ference in incomes between the new entrant and non-sugarcane households is due to sugarcane production.

From the results in the present paper, it is in- teresting to note that energy consumption levels for the new entrant group are virtually identical to sugar and non-sugar producers. Some litera- ture suggests that for crops with a long growing cycle like sugar, households may experience tem- porary shortages of food before payment for the crop is received.‘” In the present study, this does not appear to be the case.

The Kenya research is one of a series of cash cropping case studies being carried out by the In- ternational Food Policy Research Institute in collaboration with institutions and researchers in developing countries. The positive effects of sugarcane production on income, consumption, and food security in Kenya may not necessarily be replicated in each of the other country-specific studies. The extent to which this happens de- pends on factors such as effects of cash cropping on real income, sources and control of income, agricultural production strategies, expenditure patterns, household decisionmaking, demand for land and labor, and allocation of food and other resources within the family. The importance of each of these factors may vary by sociocultural environment. The conventional “wisdom” that an increased emphasis on commercial agricul- tural production necessarily results in a de- terioration of household food security is not borne out by the data from the present study.

NOTES

1. The terms “export crops” and “cash crops” are 5. Sublocation is the smallest administrative unit in often used interchangeably in the literature and this Kenya. The Sony factory serves 25 sublocations. Areas creates some confusion. Export crops are those that are closest to the factory have a higher proportion of sugar exported from the country and which can be food or farmers than areas further out from Sony. non-food crops. Cash crops are commodities sold which can also be either food or non-food crops. 6. Non-sugar households also had to: (I) have a pre-

schooler, (2) have less than 20 hectares of land, and (3) 2. Republic of Kenya (1981). be a resident farmer.

3. von Braun and Kennedy (1986). 7. Similar patterns emerge when short rains land utilization patterns are compared.

4. For a more detailed description of the sampling procedure used, see Kennedy and Cogill (1987). 8. Since the project area has both a long and short

Page 7: The commercialization of agriculture and household-level food security: The case of Southwestern Kenya

AGRICULTURE AND HOUSEHOLD-LEVEL FOOD SECURITY IN KENYA 1081

rains growing season, agricultural labor is in demand 11. Analyses of the effects on various commercial year round. For a detailed discussion of this issue, see crops including sugarcane on preschooler nutritional Kennedy and Cogill (1987). status is contained in Hitchings (1982).

9. Description and discussion of household con- 12. See Kennedy and Cogill (1987). sumption function is contained in Kennedy and Cogill (1987). 13. Examples of temporary food shortages were re-

ported for coffee in Malawi by Ogbu (1973). 10. See Fleuret and Fleuret (1980, p. 250) for a dis- cussion of examples where cash crop production has had some apparent negative effects on consumption and/or nutritional status.

REFERENCES

von Braun, Joachim, and Eileen Kennedy “Commer- cialization of subsistence agriculture: Income and nutritional effects in developing countries” Working Paper 1 (Washington, DC: International Food Policy Research Institute, 1986).

Fleuret, P., and A. Fleuret, “Nutrition, consumption, and Agricultural change,” Human Organization, Vol. 39 (March 1980).

Hitchings, Jon, “Agricultural determinants of nutri- tional status among Kenyan children with model of anthropometric and growth indicators,” PhD disser-

tation (Stanford, CA: Stanford University, 1982). Kennedy, Eileen, and Bruce Cogill, “Income and nutri-

tional effects of the commercialization of agriculture in South-western Kenya,” Research Report 63 (Washington, DC: International Food Policy Re- search Institute, 1987).

Ogbu, F., “Seasonal hunger in tropical Africa as a cul- tural phenomenon,” Africa, Vol. 46 (1973), pp. 317- 332.

Republic of Kenya, National Food Policy Paper, Sessional Paper 4 (Nairobi: 1981).