«The Role of Gender and Kinship Structure in Household Decision-Making for Agriculture and Tree Planting in Malawi Seline S. Meijera,c,*, Gudeta W. ...»
Multinomial regression analyses were used to test the association of gender and kinship with the household decision-making roles for the various agricultural activities. The model consisted of gender and kinship as control variables. This relationship was expressed as Di = f (G, K) where Di is household decision-making role (decision-making by the husband, the wife, or joint) regarding the various agricultural activities (i); and G represents the gender (male vs. female) and K is the household kinship (matrilineal vs. patrilineal). Differences in the odds of decision making roles were estimated and their significance indicated by the Wald χ2.
Chi-square tests and Kruskal-Wallis tests were used to test the association between the decision-making role for tree planting and explanatory variables (including district, age, education level, household size, size of the landholding, and membership of a farmer’s group).
Negative binomial regression analyses were used to test whether the density of planted trees was associated with the decision-making role for tree planting and tree management across kinship and gender. This relationship was expressed T = f (Dtp, Dtm) where T is the number of trees planted per hectare and Dtp and Dtm are the decision-making roles for tree planting and tree management, respectively.
The outcomes of the group exercises conducted during the focus group discussions were analyzed using descriptive statistics, including frequency tables. Chi-square tests were used to test for differences in decision-making between Mzimba and Chiradzulu and between groups of male and female respondents. Statistical analyses were supplemented by the qualitative information collected during the focus group discussions.
The data were analyzed using SPSS and SAS.
Results Household Characteristics of the Study Group Our sample included 135 married households, of which 68 were based in Chiradzulu and 67 in Mzimba (Table 1). The main ethnic groups in Chiradzulu were the Lomwe (51 percent), Yao (16 percent) and Ngoni (16 percent), whereas the most common ethnicities in Mzimba were Tumbuka (48 percent) and Ngoni (33 percent). The average household size was five people (with a Standard Deviation (SD) of ± 2) in both areas. Nearly all (99 percent) households owned land, and some respondents (21 percent) reported they rented additional land for farming. The average total farm size of the respondents was 0.63 hectares (± 0.37 ha) in Chiradzulu and 2.00 hectares (± 1.68 ha) in Mzimba, although the actual acreage under cultivation was lower in both areas.
The main food crop planted in both districts was maize, while the main cash crops were pigeon pea (Cajanus cajan) in Chiradzulu (planted by 54 percent of households) and tobacco (Nicotiana tabacum) in Mzimba (planted by 21 percent of households). Almost all households (99 percent) applied mineral fertilizer on their farm, while 62 percent of the respondents also used some form of organic fertilizer. The main source of energy for cooking was firewood (reported by 99 percent of the respondents).
Decision-Making Roles for Agricultural Activities For the various agricultural activities, most decisions were being made either by the husband or jointly by the husband and wife together (Table 2). Decision-making by the wife alone also occurred but was less common. An exception is firewood collection, which was nearly always decided upon by the wife and which is related to the fact that it is often considered the domain of women in Malawi. Tree planting and tree management appeared different from the other agricultural activities in that the percentage of cases where the husband decides independently was higher—and consequently decision-making by the wife and joint decision-making were lower—compared to the other agricultural activities. These trends were significant for all activities except the selling of farm products, in which the
proportion of decision-making by the husband, wife, and joint decision making was equal (Table 2).
Table 2: Percentage of survey respondents identifying the husband, the wife, or joint decisionmaking as the main decision-maker for various agricultural activities
The decision-making roles varied for male and female household heads as well as for matrilineal and patrilineal households (Table 3). The multinomial regression analyses showed that gender and kinship structure were not significantly associated with the decisions to plant crops, apply fertilizer, or access credit. The regression outcomes for the decisions related to other agricultural activities were significant and all outcomes are presented in Table 4. Kinship structure was significantly associated with the decision-making roles for sowing and weeding of crops, rearing animals, participating in meetings, firewood collection, tree planting, and tree management (Table 4).
Table 3: Percentage of survey respondents identifying the husband, the wife, or joint decision-making as the main decision-maker by gender of the head of the household and kinship.
For sowing and weeding of crops, rearing animals, and participating in meetings, decisions were more often made by the husband in patrilineal households, whereas matrilineal households were characterized by more joint decision-making (Table 3). For firewood collection, most decisions were made by the wife in both patrilineal and matrilineal households, but the proportion was higher in patrilineal households (Table 3).
For tree planting and tree management, decisions were more often made by the husband in patrilineal households, while the proportion of decisions made by husbands and joint decision-making were about equal in matrilineal households (Table 3). Gender of the household head was significantly associated with the decision-making role for tree planting and tree management (Table 4). For both activities, decisions were dominated by the husband in male-headed households, whereas most of the decisions in female-headed households were made by either the wife or jointly.
MEIJER ET AL -65Journal of Gender, Agriculture and Food Security Vol. 1, Issue 1, pp 54-76, 2015 Table 3 also provides interesting insights into the proportion of households where the household head is the main decision-maker. For decisions on activities such as planting, sowing, and weeding of crops, the main decision-maker is the household head in about half of the households sampled (this does not seem to differ between male- and female-headed households. -). For tree planting and tree management, however, the household head is more often the main decision-maker in male-headed households and less often in femaleheaded households compared to the other activities (Table 3). Selling farm produce also involves less decision-making by the household head in male-headed households, whereas nearly all decisions are made by the household head in female-headed households (Table 4).
Table 4: Differences in decision-making roles between genders (husbands vs. wives) and kinship (matrilineal vs. patrilineal) for various agricultural activities.
Implementation of Agricultural Activities For most agricultural activities, implementation was usually carried out jointly by the husband and wife together (Table 5).
Table 5: Percentage of survey respondents identifying the husband, the wife, or joint implementation as the main implementers of various agricultural activities
The P-values in the table represent the outcome of a chi-square test testing for a 0.33-0.33distribution of proportions.
As in decision-making, there was less joint implementation and more implementation by the husband alone for tree planting and tree management compared to other agricultural activities. Selling farm products and firewood collection are mostly done by the wife. These trends were significant for all activities except accessing credit, in which the proportion of implementation by the husband, wife, and joint decision-making was equal (Table 5).
Tree Planting and Management In this study, only a few variables were associated with the decision-making roles for tree planting within the household. A larger size of the landholding was associated with decisionmaking by the husband, whereas joint decision-makers were linked with smaller landholdings (K = 23.254, P = 0.001). Decision-making in relation to tree planting is different between the two districts, with more joint decision-making in Chiradzulu (in Mzimba it is mostly the husband who makes the decision alone) (Χ2 = 25.892, P 0.001).
There was no association between education level (Χ2 = 1.517, P = 0.824), age (K = 1.097, P = 0.578), household size (K = 1.147, P = 0.564), and membership of a farmers group (Χ2 = 4.702, P = 0.095) on the one hand and household decision-making on tree planting on the other.
The household decision-making roles regarding tree planting were associated with different outcomes with regards to actual tree planting behavior. A negative binomial regression analysis showed that both decision-making for tree planting and tree management had a significant influence on realized tree densities (Table 6). Densities were significantly (P = 0.0026) higher (136 trees per ha) when the wife makes the decision to plant trees compared to when the husband decides (34 trees per ha). However, for decisions on tree management a different pattern was found (Table 6). Densities of planted trees were significantly (P = 0.0135) lower (32.7 trees per ha) when the wife makes the decision on tree management compared to joint decision-making, which realized the highest tree density (111 trees per ha). As the 95% CI do not overlap, we can confidently state that joint decision-making on tree management is more favorable than when the wife is the decision-maker.
Table 6: Predicted tree densities (number of trees per ha) depending on who makes decisions on tree planting and tree management
Perspectives from the Focus Group Discussions The results of the focus group discussions were similar to the outcomes of the household survey. For most agricultural activities, the decision-making pattern was a mix of decisionmaking by the husband, the wife, or by both (Table 7). Decisions in relation to fertilizer application, rearing animals, and selling of farm products were more often made by the husband and wife jointly compared to other activities. However, tree planting and tree management seem to be considered mostly the domain of the husband. In contrast, firewood collection was seen as a task for the wife. Although most activities were implemented by the husband and wife jointly, again tree planting and tree management were more often implemented by the husband alone in comparison to the other activities, reinforcing the notion that they are tasks for men (Table 8). There was no difference between the groups in Mzimba and Chiradzulu, nor between the male and female groups, in the household decision-making roles; the exception was decision-making on participation in meetings, which was seen as a task for the husband in Mzimba and for the wife in Chiradzulu (Χ2 = 6.667, P = 0.036).
Table 7: Identification of decision-maker for household decisions regarding various agricultural activities
During the focus group discussions, there was a lot of debate over who makes the final decision on various agricultural activities. This shows that there was considerable variation between households in how decisions are being made and undermines the notion that there are clear patterns for the roles of the husband and wife. The focus group participants also provided more details and background on gender-specific tasks when it comes to tree planting. Although tree planting and tree management were generally seen as activities for men, there was also participation by women. Participants explained that husbands generally dig the holes, prepare the planting stations and firebreaks, and take care of the pruning and weeding, whereas the wives help with getting seedlings to the planting stations, watering the seedlings, and applying manure and sweeping the leaves around the planted trees. One participant also reported that there are gender differences for planting and managing different types of trees; for example, men generally take care of trees planted in the fields (such as fertilizer trees) whereas women look after the trees around the homestead (such as fruit trees). However, this was not mentioned during the other focus group discussions.
Discussion The results revealed that households vary in who makes agricultural decisions. It was common for the husband to be the main decision-maker or for decisions to be made jointly by the husband and the wife together. Decision-making by the wife was less common, except for firewood collection (which is considered a task for women). In patrilineal households, decision-making was dominated by the husband, whereas there was more joint decision-making in matrilineal households. Implementation of these activities was mostly done by the husband and the wife together. The results also suggest that for most agricultural activities, the household head was the primary decision-maker in about half of the households sampled. These results do not agree with the traditional belief that the household head is the chief decision-maker within rural households in Africa. In contrast, Posel (2001) examined the concept of headship for self-reported household heads in South Africa and found that the concept is still valid, as heads are in fact the key decision makers in the households studied. However, our results are more in line with recent findings by Mbweza et al. (2008), who examined the decision-making process of husbands and wives in matrilineal and patrilineal families in Malawi covering various areas of decisions (including money, food, family planning, and sexual relations). They found that most couples used a mix of decision-making approaches. These conclusions reinforce our finding that the household head is not always the chief decision-maker and show that alternative decisionmaking approaches are also being employed.