The influence of smallholder labour demand on cultivation of cash crops in northern Mozambique

Abstract

Labour is one of the most important factors affecting smallholder cultivation of cash crops. Available household labour (AHL), crop labour requirements (CLR) and the ratio AHL:CLR were analysed from data collected from 287 households in the southern Niassa province of Mozambique. The study confirms that, other factors being held constant, crop labour requirements were positively related to the number of households rejecting or discontinuing certain cash crops owing to lack of available labour. Weeding was the most labour-intensive operation, followed by harvesting, preparing soil, transporting produce, clearing land and preparing seedlings. The following labour-dependent factors can be estimated: (i) the total area a household can cultivate, (ii) the area that can be allocated to food crops for consumption, (iii) the area that can be allocated to cash crops, (iv) the proportion of households that can cultivate cash crops, and (v) the proportion of households that are unlikely to produce sufficient food for household consumption.

¹Respectively, Country FEWS NET Representative, Angola Ministerio de Agricultura e Desenvolvimento Rural, Luanda, Angola; Professor and Head of Discipline, Community Resources, School of Agricultural Sciences and Agribusiness, University of KwaZulu-Natal, Pietermaritzburg; and Director (Teaching and Learning) and Academic Director (Extension & Resource Management), Centre for Environment, Agriculture and Development, School of Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg.

1. Introduction

This study is part of an overall research project on factors that affect smallholder cultivation of cash crops in northern Mozambique. Two earlier studies identified labour as one of the most important factors influencing smallholders' decision to cultivate cash crops. Lukanu et al. (2004) identified, in order of importance: (i) profitability, (ii) market access, (iii) availability of inputs and implements, (iv) labour, and (v) access to extension services; and Green et al. (2006) reported that the number of effective labourers was the most determining factor that explained variation in market participation. Makhura (2001) also found that labour factors influenced smallholders' participation in agricultural markets. The amount of production depends on the size of the field that can be cultivated, which in turn depends on the availability of labour (Larson & Frisvold, 1996). Insufficient labour can limit the ability to sow a large area quickly (Abalu et al., 1987), and inadequate production, because the area of cultivated land is reduced, limits market participation.

Green et al. (2006) reported that the majority of farmers in the southern Niassa province of Mozambique gave priority to cultivating food crops for consumption: their available household labour was first assigned to these and what was left over was used to cultivate cash crops. Enete et al. (2002) analysed the contribution of labour by men and women to food crop production in Africa and found that the number of households where females provided more field labour than males was higher for female-headed than male-headed ones. However, there is a need for more elaborate research into the socio-economic aspects of labour and its physical complexity. This will lead to greater understanding and proper strategies for helping smallholders cultivate cash crops. This study investigated how the independent and combined aspects of available household labour and crop labour requirements affected households' involvement in cultivating cash crops and their subsequent market participation in southern Niassa. Its specific objectives were to analyse the relationships between:

(i) available household labour (AHL) and smallholder cultivation of cash crops (on the hypothesis that available household labour would be positively related to the number of cash crops cultivated by the household);

(ii) the crop labour requirements (CLR) and smallholder cultivation of cash crops (on the hypothesis that, other factors being held constant, crop labour requirements would be positively related to the proportion of smallholders not cultivating a given cash crop owing to lack of available labour); and

(iii) the ratio AHL:CLR and cultivation of cash crops (on the hypothesis that the ratio AHL:CLR, which represented the potential field size, would be positively related to the number of cash crops cultivated).

2. Analytical Methods

The study analysed three labour factors that were likely to influence smallholder cultivation of cash crops: available household labour (AHL), crop labour requirements (CLR) and the ratio AHL:CLR. From the literature, the authors recognise that there are a myriad other factors that also influence smallholder cultivation of cash crops, such as those shown in Figure 1. Makhura (2001) investigated the most important factors influencing participation in agricultural markets in Limpopo, South Africa. Research by Matungul (2002) analysed the marketing constraints faced by communal farmers in KwaZulu-Natal, South Africa. This paper analyses only the labour aspects that influence smallholder cultivation of cash crops in southern Niassa, Mozambique. It uses some equations that combine socio-economic and agronomic data to demonstrate and explain how labour affects the cultivation of cash crops. These equations are discussed below.

Figure 1: Factors and related interactions affecting smallholder cultivation of cash crops and participation in agricultural markets (Lukanu, 2005)

2.1 Available household labour

Household labour capacity or available household labour is a measure of the household's total physical labour capacity (Sharp, 2003). AHL represents the number of person-days a household has for household farm labour and can be estimated using the following equation:

The study used the approach by Green et al. (2006) based on effective household labour to estimate the number of persons. EHL was calculated by summing active household members (AHM), determined from the respondents' own assessment, with hired external labourers (EL) minus active household members who provide labour to other households (HMEL) (Green et al., 2006).

Other approaches have also been used to estimate the number of persons involved in farm labour. For example, Owuor (n.d.) estimated the number of persons involved in farm labour by taking into consideration correction factors that converted non-adult members of various age categories to adult equivalents and a study by Ravallion (1992) discussed some problems related to the process of estimating the adult equivalent.

Days means the number of days a household has to perform various livelihood activities. In this study the peak period was used. This is because labour during peak period determined the cultivated area (Jeje et al., 1998). The number of days was determined by multiplying the number of weeks during peak period (Week) and the number of working days per week (Dweek).

For example, the peak period of about 16 weeks between mid-October and mid-February multiplied by 6 working days/week equals 96 working days (Days = Week x Dweek = 16 × 6). Boeteng et al.'s (1987) definition of a man-day is 8 hours/day per person, which is similar to the Mozambican standard day (Sday) (Labour Ministry Officer in Cuamba district, 2002, personal communication). The correction can be performed by multiplying 96 by the ratio of actual hours/days (Hday) to the standard day (Sday = 8 hour/day). Equation 3 then becomes

Hday:Sday is referred to as the correction factor. For example, on average smallholders worked for 7.2 hours/day (Lukanu, 2005). The corrected days can then be calculated using Equation 4 (16 × 6 × 7.2/8 = 86.4 days). That is, the 96 days of 7.2 hours/day equal 86.4 days of 8 hours/day. The left-hand and right-hand terms of this equation correspond to the same number of hours worked (96 days × 7.2 hours/day = 86.4 days × 8 hours/day = 691.2 hours during peak period). A more detailed estimation could account for days when people did not work because of holidays, funerals or illness.

2.2 Crop labour requirements

Crop labour requirements are defined in this study as the number of days (Day) a person (Person) takes to cultivate an area of land of one hectare (Ha) for a given crop. This can be written as follows:

These requirements can be calculated by first determining the labour requirements of various agricultural operations (LRAO) then summing these to calculate the CLR for a certain crop. Therefore,

Persons in Equation 6 means the number of people involved in performing an agricultural operation on one hectare of land. For example, an average household in southern Niassa requires two people to perform the first and second weeding operation and 1.8 for threshing operations (Lukanu, 2005). Enete et al. (2002) collected information using a rapid rural appraisal technique where groups of men and women were interviewed about time use and the number of people involved in agricultural operations.

Days in Equation 6 means the number of days taken to perform an agricultural operation. There was also a need to determine the number of hours per day that people worked (Hday), which varied with operations. For example, smallholders in southern Niassa took on average 6.8 hours/days to perform the first weeding and 4.6 hours/days to perform the second weeding (Lukanu, 2005). Taking the above into consideration, days worked for a given operation can be calculated as follows:

2.3 Potential area cultivated

Available household labour and crop requirements during peak period will determine the area a household can cultivate or weed. Jeje et al. (1998) point out that at peak times labour such as weeding makes it difficult for a household to increase its cultivated area. The greater the available household labour, the greater the area a household can cultivate or weed during the peak period. Or, put differently, the greater the crop labour requirements, the smaller the area it can cultivate or weed during this period. That is, the cultivated area (A) is directly related to household available labour and inversely related to crop labour requirements. This can be expressed as follows:

By substituting the meanings of AHL and CLR (as shown in Equations 1 and 5) it can be demonstrated that AHL:CLR expresses the size of the cultivated field in hectares (A = Persons x Days:Persons x Days/Ha). Equation 8 assumes that households use most of their available labour during the peak period of agricultural activities. In addition, the equations above are valid only for an area where human power is the main source of energy, as it is in southern Niassa.

The ratio of AHL to CLR is likely to provide a useful explanation of the way labour demand limits the area that can be cultivated for food and cash crops. Many studies on market participation have shown that the size of the cultivated area determines who can cultivate a cash crop and subsequently participate in agricultural markets (Abalu et al., 1987; Larson & Frisvold, 1996; Makhura, 2001).

2.4 Area for food crop for consumption and respective AHL and CLR

The area for cultivating a food crop for consumption (Afood) can be determined using data about households' consumption requirements (HHcomsup) and yield (Yield). Maize is the chief staple food crop in southern Niassa. The area required to cultivate maize for household consumption can be estimated as follows:

The household consumption requirements for maize can be estimated by multiplying the total number of household adult equivalents by the amount of maize consumed per adult equivalent. According to Akesson (1994), an adult member in Mozambique (Tete Province) consumes 150 kg of maize per year. An equation used by Haarmann (2000) can be used to estimate the total household adult equivalent for consumption – Total household Adult Equivalent = (Adult + 0.5 × Children)^0.9. Knowing the size of the area assigned to cultivate food crops for consumption is necessary because it gives an idea who can cultivate cash crops. According to Green et al. (2006), households in southern Niassa give priority to food crop cultivation before planning to cultivate cash crops.

It is possible to determine the amount of household labour required to cultivate food crops for consumption during the peak period using Equation 8 and the area required to cultivate a food crop for consumption, as follows:

This equation indicates which households are likely to experience a food shortage and are therefore less likely to cultivate cash crops. This will happen when the required labour to cultivate food crops for consumption during the peak period (AHLfood) is larger than the household's available labour. In this situation, the household is unlikely to cultivate an area large enough to produce sufficient food for its consumption and less likely to venture into cultivating cash crops.

2.5 Area for cash crop

Households can cultivate a cash crop only if they have some labour available after securing the labour required to cultivate a food crop for consumption. That is, the available household labour for a cash crop during the peak period (AHLcash) can be estimated as the difference between the total AHL and the portion used for cultivating a food crop for consumption (AHLfood):

Applying Equation 8, the area for a cash crop of known labour requirements can be estimated using the following equation:

This equation gives the area of land the household would use to cultivate a cash crop if it used all the labour it has available after deducting the labour required to cultivate the food crop for consumption.

2.6 Statistical analysis

The above equations were used to estimate the following factors that influence the cultivation of cash crops: (i) available household labour, (ii) crop labour requirements, (iii) total potential area, (iv) area for food crop for consumption and respective AHL and CLR, and (v) area for cash crop. The relationships between these factors and the number of cash crops cultivated were analysed using (i) descriptive statistics, and (ii) correlation at α = 0.05 to find the direction (positive or negative) and whether the relationships were significant or not. It was assumed that farmers who diversified into more cash crops were likely to be classified as better producers of cash crops, ¹ because crop diversification serves as a central strategy in managing risk (Makhura et al., 1997; Anderson, 2003). Diversification also increases productivity and the subsequent cash that a household can earn from selling crops (Anderson, 2003). An attempt was also made to compare the areas estimated in the equations above with the areas estimated in the literature on southern Niassa, reviewed above.

3. Aspects of Southern Niassa Cropping System that Affect Labour

Southern Niassa has a population density of 13 people per sq km, so availability of land is not the factor limiting cultivation (Instituto Nacional de Estatística, 1997). According to the Programa de Apoio aos Mercados Agricolas (PAMA, 2003), the factors that limit land to only about 2.45 ha per household are insufficient effective household labour and the use of inefficient manual agricultural implements such as hoes, machetes, axes and sickles. The average effective household labour in southern Niassa is 2.3 persons (Green et al., 2006). Smallholders use their labour resources firstly to cultivate food crops (maize, sorghum, cassava, boer bean, cowpea and bambara nut) and only secondly to cultivate cash crops (food cash crops, i.e. household surplus, and promoted cash crops such as tobacco, cotton, sesame, sunflower and paprika).

Casual labour is scarce during the peak period that coincides with the single rainy season between November and April. Household members prefer to cultivate their own fields during this period to guarantee household consumption. Only poor farmers and migrant labourers from neighbouring districts are available for hire. The neighbouring Zambezia and Nampula provinces are sources of labour for Cuamba district because the bigger economic markets in Cuamba ² provide a greater attraction for labour. Wealthy farmers can hire casual labour, but the majority of poor households are too short of income and food to be able to afford it.

4. Methodology

4.1 Sampling research tools

A detailed description of this study's methodology can be found in Lukanu (2005). The research combined cluster and stratified samplings. The random selection of one or two villages for each of the six routes in Cuamba district ensured that the study's results could be generalised for Cuamba and the neighbouring districts. Individual households were identified by wealth-ranking criteria and subsequently interviewed. Nine focus group discussions involving key informants such as traditional, church, mosque, political and government authorities, and other opinion leaders, preceded the study. Information elicited from these groups served as the basis for the design of questionnaires. A total of 287 household heads were selected to respond to a set of questions. In cases of absence, their spouses responded or consulted the members of the household for answers to questions they were uncertain about.

4.2 Data collected

Table 1 shows the data used in this study and how they were collected. These data can be categorised as available household labour, crop labour requirements, household consumption requirements, crop yields, wealth status, and strategies suggested by households for improving labour use for agricultural activities.

Table 1: Data used in this study and how they were assessed (ranks used: 0 = lowest rank and 5 = highest rank)

Data	How estimated
Active members of household (Ham)	By asking respondents how many household members were actively involved in agriculture. All active members were considered as adults.
External labour or ganho-ganho (El)	By asking respondents how many ganho-ganho the household employed during peak period.
Household members providing ganho-ganho to others people's farms (HmEl)	By asking respondents how many household members provided ganho-ganho to other people's farms.
Working hours per day (Hday)	Through various group discussions on time use in agriculture and for each agricultural operation, such as departure time, arrival time, work time, rest time and leaving time from the field.
Persons involved per operation	By asking respondents how many people were involved in performing each agricultural operation.
Days taken to perform an operation, in hectares of land	By asking respondents to recall number of months (month), number of weeks per month (Wmonth), and number of days per week (Dweek) taken to perform a given operation.
Household consumption	By multiplying number of household members by 150 kg of maize.^1
Yield	By using yield from literature review about northern Mozambique agriculture.
Number of promoted cash crops sold	By asking respondents how many cash crops they had sold during the last agricultural market campaign.
Perceived labour-intensive months, operations and crops	By asking respondents what they thought were the most/least labour-intensive months, operations and crops.
Household wealth status	Through key informants and use of visual indicator of poverty.
Farmers' suggested strategies for alleviating labour constraints	By asking respondents what strategies that they thought could help alleviate labour constraints.
^1Household adult equivalent was not estimated because the data did not allow such a calculation.

5. Results and Discussion

5.1 Available household labour (AHL)

On average, households worked about 7.2 hours per day and six days per week, except during peak periods when some worked for seven days per week. On average 2.3 people were involved in the household's fields, including household members and external labour. The average annual available labour was 652 person-days/year and 200 person-days for the 16-week peak period.

It was hypothesised that AHL can be positively related to the number of cash crops cultivated by a household. The AHL during the peak period (Equation 1) was positively and significantly (R = 0.179; α = 0.002) related to the number of promoted cash crops sold (PCROSOLD) (see Table 2). Households with less labour were less likely to cultivate cash crops as they would use most of their labour to cultivate food crops for consumption. Equation 2 also suggests that households can increase their EHL (external hired labour) by employing external labour. Carrilho et al. (2003) reported that approximately 20 per cent of Mozambican smallholders employ casual labourers (known locally as ganho-ganho) to work in their fields. Active household members were more likely to cultivate their own household's fields than those of others. However, AHL during the peak period explained only 3.2 per cent of the variation in the number of promoted cash crops cultivated. Other labour-related factors analysed below and some other factors shown in Figure 1 may have influenced smallholders' cultivation of cash crops.

Table 2: Statistics of a regression model between available household labour (AHL) during peak period and the number of promoted cash crops cultivated (n = 287, Cuamba, September 2002)

R	R square	Adjusted R square	Std. error of the estimate	Change statistics					Durbin-Watson
				R Square change	F Change	df1	df2	Sig. F change
.179^ab	.032	.029	.746	.032	9.399	1	284	.002	1.605
^aPredictors: (Constant), AHL.
^bDependent variable: PCROSOLD = Promoted cash crops sold.

The calculated average AHL for food crops for consumption (Equation 10) and the remaining available labour (Equation 11) for cash crops were 104 person-days and 97 person-days during peak period, respectively. Households that use all their AHL for their own food crops for consumption (i.e. 1:1) have nothing left over for cultivating cash crops. But if they use only part of their AHL for this purpose then they will have some left over for cultivating cash crops. Put another way, those who have more AHL than they need for their own food crops are more likely to be able to cultivate cash crops. Figure 2 shows the distribution of ratios of AHL to labour used for cultivating food crops for consumption for the 287 households surveyed in this study. Those on or below the 1∶1 line are the ones more likely to cultivate a cash crop. Those above the line are unlikely to be able to, even if they want to, because of the labour limitation. The Ministério de Agriculture e Pescas (1996) reported that the strategies of households in Cabo Delgado and Nampula provinces emphasised firstly food security, but they were open to cash crops when there was sufficient available labour.

Figure 2: Ratio of total available household labour (AHL) to labour allocated to food crops for consumption (n = 287, Cuamba, September 2002)

The estimates suggested that 32 per cent of the sample (n = 287) – 40 per cent of the poor, 28 per cent of the middle-class and 14 per cent of the wealthy households –would have insufficient labour to cultivate a big enough field to produce cash crops. The survey identified 26 per cent of the sample – 37 per cent of the poor, 17 per cent of the middle-class and 17 per cent of the wealthy households – that did not sell any cash crops. Van Zyl and Coetzee (1990) reported data for Zimbabwe (15–26 per cent) that were similar to this estimate, showing that the proportion of deficit producers can double in a year of low rainfall owing to poor yields. These households were then less likely to cultivate cash crops because their main goal was to produce food for consumption. They reported that such producers were also less likely to respond to price incentives because they had insufficient labour to cultivate cash crops.

It is important to stress that households have an untapped labour availability during the dry season that could be used for cultivating cash crops. More than 69 per cent of this labour was available during the non-peak period, mainly during the dry season. Part of it was used to perform some agricultural operations (not including the operations shown in bold in Table 3, such as tree felling, land clearing and soil preparation). Other available labour was used for non-agricultural activities such as hunting, building and maintaining houses. Overall, households under-utilised their available labour during dry seasons compared to the peak period.

Table 3: Monthly ranking of labour demand and agricultural operations throughout the year in southern Niassa

The cultivation of cash crops could be increased by promoting dry season cash crops to help households make good use of their labour during this season. Shifting the cash crop labour from the rainy to the dry season could reduce the competition for labour between cash and food crops and thus increase households' annual outputs and income. Cultivating during the dry season can be performed under small-scale irrigation schemes or on lowland ³ that holds enough soil water to sustain crop growth. However, a feasibility analysis is needed to determine the possibilities. For example, the importance of smallholders' current occupations during the dry season could be analysed in relation to cash crop cultivation.

It can be concluded that AHL during the peak period affects the smallholders' decision to cultivate cash crops. All the households with available labour less than the amount required to cultivate food crops for consumption were less likely to cultivate cash crops, even if they wanted to, because of the labour limitation. To increase the cultivation of cash crops, it is recommended that households increase the amount of AHL and use it more efficiently, for example by cultivating cash crops during the dry season.

5.2 Labour demand and crop labour requirements (CLR)

Table 3 and Figure 3 show farmers' monthly ranking of labour requirements in the agricultural calendar. The months between March and September were considered the least labour-intensive (ranking less than 5.5) and those between October and February the most. Other factors being held constant, farmers would be able to participate actively in the cultivation of cash crops during the less labour-intensive months that coincide with the dry season, if supported to cultivate in lowlands and/or under irrigation.

Figure 3: Monthly ranking of labour demand and agricultural operations throughout the year in southern Niassa (based on Table 3)

Nevertheless, promoters often recommend that their cash crops be cultivated between October and February, in the rainy season. Smallholders are likely to reject this recommendation because this is when they prioritise food crops for their own use. Cash crops would create competition for the available labour.

Table 3 also shows the various agricultural operations that a household can perform: clearing the land as early as March, harvesting the produce as late as September the following year, and threshing, grading, transporting and selling (post-harvest operations) the following December. The most critical and intensive activities are performed in the labour-intensive months between October and February (Table 3, operations in italics).

5.2.1 The estimated CLR

Table 4 shows the labour requirements for cultivating food and cash crops: the hours needed and the number of people involved in the various operations. The authors acknowledge the difficulty of estimating these requirements on the basis of farmers' recall. This was a validation exercise asking farmers how much time various agricultural operations took and to rank the demand for labour required by various crops (Table 5). There was positive (R = 0.731) and significant (α = 0.049) correlation between the crop labour requirements calculated using Equation 5 to Equation 7 and farmers' ranking of labour demand. The significance would have been greater if farmers had not under-ranked the labour demand for food crops because of their greater experience and the importance of food crops.

Table 4: Labour requirements for cultivating food and cash crops

	Hour/ day	Correc-tion factors^1	Average number of people involved	Food crops^2	Tobacco	Cotton	Sesame^3	Sunflower^3	Paprika	Farmers indicating labour (%)
Land clearing/ tree felling	5.8	0.7	1.5	32	32	32			32	7
Seedling preparation	4.4	0.6	1.3		8				8	5
Ridge making^4	4.3	0.5	1.8		21				21
Soil preparation	4.3	0.5	2.0	30	30	30			30	9
Sowing of early crops	4.3	0.5	2.4	12		12
Sowing of late crops	3.3	0.4	2.4	12			12	12
Transplanting	4.4	0.6	2.4		10				10
Fertiliser application	2.4	0.3	1.0		3	3			3
First weeding	6.8	0.9	2.0	41	24	40			36	43
Second weeding	4.6	0.6	2.0	22	15	22	15	15	21
Third weeding	4	0.5	2.4			14	14
Thinning	3.5	0.4	2.2			17	8	9
De-budding	3.5	0.4	1.8		7
Pest and disease control	2.9	0.4	1.0		4	4			4
Harvest	4.2	0.5	2.1	24	71	35	17	19	13	19
Threshing	1.8	0.2	1.8	18			3	2
Dryer preparation	5.4	0.7	1.5		20				11	5
Drying of products	7.7	1.0	1.5		18				15	3
Transport to the village	3.4	0.4	1.8	17	46	25	1		10	9
Grading	5.3	0.7	1.9		23	15			15
Bagging and storing	4.2	0.5	1.0	3	3	3	3	3	3
Transport to the market	3.8	0.5	1.5	2	2	2	1	1	2
Total person-days/ (ha x year)				213	337	254	74	61	234
Total person-days/ (ha x [Oct-Feb])				87	226	112	49	36	74
Note: Italics indicate the most critical and intensive operations (n = 16, Cuamba, October 2002).
^1Correction factor obtained by dividing the number of days worked by 8 hours per days, the standard working time for Mozambique.
^2Cash food crops include maize, cassava, sorghum, boer bean and cowpea.
^3Sesame and sunflower are often intercropped in food crop fields; therefore there is no need for land clearing, soil preparation and first weeding.
^4Making ridges for planting.

Table 5: Correlation between calculated crop labour requirements (CLR) and the ranked labour demand by smallholders (1 = least labour-intensive while 6 = most labour-intensive, n = 6, Cuamba, October 2002)

	CLR (person-days/[ha/ year])	Labour demand ranking^1
Food crops	213	1
Tobacco	337	6
Cotton	254	5
Sesame	74	3
Sunflower	61	1
Paprika	234	4
Person correlation		0.731
α		0.049
^1The numbers suggest that labour was the most important factor for tobacco and cotton while foodcrops and sunflower were the least labour intensive.

5.2.2 CLR and cash crop cultivation

It was hypothesised that, other factors being held constant, labour requirements would be positively related to the proportion of smallholders not cultivating a given cash crop owing to lack of available labour. Tobacco cultivation was the most labour-intensive of all of the cash crops, followed by cotton, paprika, food crops, sunflower and sesame. The percentage of households not cultivating a given cash crop owing to labour constraints was significantly and positively related to labour demand as perceived by smallholders (R = 0.826, α = 0.021) (Table 6). Therefore, other factors being held constant, smallholders would decide not to cultivate those cash crops that had high labour demand.

Table 6: Farmers not cultivating owing to lack of available labour, the estimated crop labour requirements and the farmers' ranked labour demand (n = 6, Cuamba, September 2002)

	Not cultivating (%; n = 287)	Not cultivating owing to lack of available labour (%)	Crop labour requirements (person-days/ [ha year])	Labour demand rank	Importance of labour in relation to other factors^1
Tobacco	60	56	337	6	1
Cotton	77	41	254	5	1
Sesame	93	24	74	3	2
Sunflower	95	28	61	1	2
Paprika	99	24	234	4	2
Food crops	0	0	213	1
Correlation coefficient			0.451	0.826
α			0.185	0.021
^1These numbers suggest that labour was the most important factor for tobacco and cotton but second in importance for sesame, sunflower and paprika.

However, a positive but non-significant correlation was found between the percentage of farmers not cultivating a specific cash crop and the calculated crop labour requirements (R = 0.451, α = 0.185). This poor correlation resulted because the calculated crop labour requirements, based on time, did not take into consideration the smallholders' experience, the importance of food crops and the emotional and psychological (non-physical) aspects of labour that influenced smallholders' decision to cultivate a cash crop. For example, farmers considered preparing tobacco seedlings to be more labour demanding (8 person-days/ha) than preparing tobacco dryers (20 person-days/ha) or transporting tobacco from the field to the village (46 person-days/ha). The skill and discipline required to prepare healthy seedlings and the difficulty of finding a secure source of water for irrigation were two of the factors that made smallholders feel that seedling preparation was more labour demanding than the above-mentioned activities that took more time. That is, other factors being held constant, smallholders could cultivate crops that require simple skills rather than those that require more complex ones.

Table 6 also shows that labour was the primary factor influencing farmers' rejection or discontinuation of tobacco and cotton, while it was a secondary factor influencing sunflower, sesame and paprika cultivation. Nevertheless, there were more smallholders marketing tobacco (40 per cent) in spite of its having high labour requirements or being ranked as the most labour demanding crop. The profitability of the crop, input support, technical assistance and good services for buying tobacco made farmers choose to cultivate and market tobacco rather than sunflower and sesame, which had lower labour requirements per hectare.

5.2.3 Labour requirements for various agricultural operations

An attempt was made to identify the most labour-intensive operations based on estimated labour requirements and farmers' ranking of the labour demands for various agricultural operations on a hectare of land. According to farmers, weeding was the most intensive, followed by harvesting, preparing soil, transporting produce, clearing land and preparing seedlings. Farmers' ranking of operations in terms of labour demand did not correlate significantly (R = 0.301, critical r = 0.700) with the estimated labour requirements of the operations, using Equations 5 to 7. This is because when ranking the labour demand for agricultural operations farmers took into consideration not just the time required but also the intensity, pain and complexity of these operations and the experience required.

(a) Weeding

A considerable percentage (43 per cent) of the farmers said weeding was the most intensive operation. They worked under stress during weeding because weed growth overtook their capacity to keep the weeds down. They were unlikely to clear and prepare more land during the less labour-intensive months when they knew that they would only be able to weed 2.45 ha because of serious labour shortages at this time (Chatizwa & Vorage, 2000; PAMA, 2003; Mganilwa et al., 2003). This suggests that, other factors being held constant, they were less likely to cultivate the cash crops that require intensive weeding.

On average, farmers required two weedings for most crops and three for cotton. The total labour requirements for weeding were highest for cotton (76 person-days/ha), followed by food crops (63 person-days/ha), paprika (57 person-days/ha), tobacco (39 person-days/ha), sesame (29 person-days/ha) and sunflower (63 person-days/ha). Weeding demanded about 27 per cent of the agricultural labour in southern Niassa. Riches et al. (1997) indicated that weeding accounted for up to 60 per cent of labour in maize cultivation in Zimbabwe. The intensity of the weeding operation could also be observed by the number of working hours per day (6.8) and days per week (7) it occupied (Table 4). Farmers also reported that the body position for weeding was painful – a function of low-level technology (hoes) in southern Niassa.

The period of the first weeding is characterised by heavy rain that could last ten or more days without stopping. Smallholders said this caused delays and made the weeds grow rapidly. The first two weeding operations were performed in the hungriest period (October to February; see Table 4) when farmers' energy levels were lowest because food was scarce. During this period farmers were also prone to diseases (cholera and malaria) because of the increased spread of water-borne micro-organisms and low body resistance caused by poor nutrition. Some farmers suggested that they would like to hire or buy tractors or motorised cultivators on credit to reduce the labour required for weeding.

(b) Harvesting

Harvesting was cited (by 19 per cent of respondents) as the second most labour-intensive operation. It was most intensive for tobacco (71 person-days/ha), with cotton (35 person-days/ha), food crops (24 person-days/ha), sunflower (19 person-days/ha), sesame (17 person-days/ha) and paprika (13 person-days/ha) following. Five per cent of farmers said they needed to hire ganho-ganho for harvesting. They also pointed out that the body position (similar to weeding) required for harvesting some ground produce such as cowpeas and bambara nuts also caused pain. They said some crops, such as bean and sesame, needed to be harvested early in the morning and in the shortest possible period to avoid losses. This caused farmers to cultivate just the amount they could harvest, thus limiting the amount they could sell. It is therefore recommended to promote crops which involve less time-stressed limits for harvesting and thus allow smallholders to delay harvesting for periods when other labour requirements were lower. For example, maize, paprika and cassava can be harvested when smallholders are ready, unlike crops such as sesame, boer bean and cowpea that have to be harvested as soon as they are ready in order to avoid deterioration and losses.

(c) Soil preparation

Farmers pointed to soil preparation as one of the very intensive operations. Thirty person-days/ha were required to prepare the soil using hoes. More than 17 per cent of the respondents said they needed ganho-ganho to prepare soil. Some said they would like to hire or buy tractors or motorised cultivators on credit to reduce the labour required for preparing soil. As with weeding, farmers said manual soil preparation was very strenuous given the bodily position and the energy required for removing grass and roots and ploughing to a reported depth of 0.2 m.

(d) Transport of produce from the field to the village

Similarly, farmers identified transporting their produce from the fields to the villages and the markets as labour-intensive. They made many trips on foot, carrying loads on their heads, or by bicycle. Tobacco took the most time to transport (48 person-days/ha), compared to cotton (25 person-days/ha), food crops (17 person-days/ha), paprika (10 person-days/ha), sunflower (2 person-days/ha) and sesame (1 person-day/ha). The Niassa government recommends that villages should be located close to the main roads where people would have access to transport. However, there is a need to establish buying points closer to more villages, using local buyers, farmers' associations or traders. This would reduce the time taken to transport goods to marketplaces.

(e) Land clearing

Tree-felling and land clearing were mentioned as the next most labour-intensive operation. This operation required an average of 32 person-days/ha. Although painful, tree-felling operations were rarely performed under stress as farmers could start this as early as March and finish in September (Table 3). A small proportion of the farmers (6 per cent) said they needed to hire ganho-ganho to fell trees. This was exclusively men's work so female-headed households had to pay for ganho-ganho to do it. In general, households prefer to use land that has already been cleared to cultivate promoted cash crops than to clear a new field.

(f) Seedling preparation

Tobacco and paprika growers pointed out that seedling preparation was a very complex operation. However, it required only eight person-days/ha. They had to water seedlings twice daily during the dry period. Experience from NGOs promoting paprika indicated that when farmers lost seeds and seedlings it was mostly through poor preparation. Some plants died from lack of water or poor pest control, and some seeds and seedlings were washed away by floodwater because they were set too close to the river banks. A possible strategy for reducing the amount and complexity of the labour required for preparing seedlings would be for the promoters to prepare the seedlings they sell to inexperienced farmers, or supervise the preparations until the farmers had learnt to prepare them by themselves.

5.2.4 Summary

Helping smallholders to reduce the labour load of agricultural operations, mostly weeding, during the peak period would encourage them to cultivate cash crops. Mechanised technology would reduce the CLR: only 1.4 person-days would be needed to weed a hectare, which is 20 times faster than hand-hoe weeding (Mganilwa et al., 2003). However, a feasibility study would be needed to assess the skills, maintenance, management knowledge and initial and operating expenses of mechanisation in such rural areas (Haque et al., 2000). The other most economical and appropriate source of energy not mentioned by the respondents is draught power (Devendra & Thomas, 2002). However, much work would need to be done to introduce draught technology to southern Niassa. Intercropping, already a part of the cropping system in southern Niassa, can also help smallholders reduce the CLR. For example, sesame or sunflower intercropped with maize would not require any additional labour in terms of land clearing and soil preparation.

5.3 Land areas for cultivation

Equation 8 was used to estimate the area a household could cultivate based on its AHL and CLR for food crops. Ideally, the study should measure the total field size a household cultivates and compare it with the estimated data to test the accuracy of Equation 8. The estimate suggests that on average a household can cultivate 2.3 ha of land if it cultivates only food crops. This estimate is very close to the average area of 2.45 ha distributed in 2.5 fields per household for Cuamba district (PAMA, 2003).

The estimated average field size required to cultivate food crops using Equation 10 is 1.2 ha. ⁴ This field size is close to PAMA's (2003) measured area of 1.22 ha per household to cultivate food crops for consumption. That is, a typical household first needs to assign the AHL required to cultivate 1.2 ha for food crops for consumption, while the remaining, if any, AHL can be used to cultivate cash crops. On the basis of Equation 10, the area a smallholder assigns to cultivating food crops for consumption could be reduced and that for cash crops increased if the smallholder was helped to increase the yield of food crops with the application of fertilisers and/or good agricultural practices. By applying this strategy, smallholders could be helped to reduce the amount of AHL needed to cultivate food crops for consumption and thus have some spare for cultivating cash crops.

The estimates using Equation 12 also suggest that if a household used all the remaining AHL for cash crops it could cultivate 0.4 ha of tobacco, 0.9 ha of cotton and 1.3 ha of paprika. These areas vary because of differences in the CLR: a smaller field size would be cultivated for a cash crop with a higher CLR. The correlation between the estimated area and the number of cash crops sold was positive and significant (R = 0.138, critical α = 0.020). That is, the ratio AHL:CLR, or the estimated land area was positively related to the number of cash crops cultivated (Table 7).

Table 7: Regression model between estimated area for cash crop and number of promoted cash crops sold by smallholders farmers (n = 287, Cuamba, September 2002)

R	R square	Adjusted R square	Std. error of the estimate	Change statistics					Durbin- Watson
				R Square change	F change	df1	df2	Sig. F change
				.138^a	.019	.016	.750	.019		5.502	1	284	.020	1.602
^aPredictors: (Constant), AREACASH (Land area under cash crop).
^bDependent variable: PCROSOLD (Number of promoted cash crops sold).

Estimated data of potential field size for various cash crops were compared with the assessed data from literature and information from promoting institutions to assess the validity of the concepts and equations developed in this study. PAMA's (2003) assessed land size of one hectare for cotton is very close to the 0.9 ha estimated in this study. According to the João Ferreira dos Santos Company senior staff, each farmer cultivates on average about 0.3 ha of tobacco in southern Niassa. These data are also close to the estimated average of 0.4 ha for tobacco in this study. There were no independent assessments of cultivated areas for sesame and sunflower for comparison. It is important to stress that on average households used their available labour to cultivate 1.77 ha for food cash crops and/or promoted cash crops. The suggestion that a household would use all its remaining available labour to cultivate a single cash crop is hypothetical, just for analytical purposes.

The combination of available household labour and crop labour requirements during the peak period therefore determined the size of the land that a household could cultivate. It would only clear a new field that it would be able to cultivate during the peak period (mainly weeding). The size of the field would dictate how much land a household could assign to cash crops after deducting the amount of land required for food consumption.

The assumptions developed in this paper can be used to estimate the size of the field and the proportion of households likely to cultivate cash crops, using simple equations involving socio-economic and agronomic data. These data include household size, maize consumption per adult person per year, total available household labour during the peak period, crop labour requirements for food crops, promoted cash crops during the peak periods, and yield.

6. Conclusions

This study investigated three labour concepts that influence the cultivation of cash crops: available household labour (AHL), crop labour requirements (CLR), and the ratio AHL:CLR that represents the field size. Available household labour and the ratio AHL:CLR for peak periods were positively related to household involvement in the cultivation of cash crops. Crop labour requirements were positively related to the proportion of households not cultivating a given cash crop owing to lack of available labour. Households first use their available labour to cultivate food crops for consumption and the remainder is used to cultivate cash crops (food cash crops and promoted cash crops). Understanding these concepts has implications for the formulation of strategies to alleviate labour constraints in order to encourage smallholders to cultivate cash crops and participate in the agricultural market.

Using these three concepts, the study was also able to estimate the potential area that a household can cultivate, the areas likely to be allocated to food crops for consumption and to cash crops, the proportion of households that could cultivate cash crops and the proportion that were unlikely to produce sufficient food to meet household consumption requirements. The estimates from this study compared well with independently surveyed data. However, further study is required to obtain empirical evidence to test and improve the accuracy of these estimates.

Notes

¹Respectively, Country FEWS NET Representative, Angola Ministerio de Agricultura e Desenvolvimento Rural, Luanda, Angola; Professor and Head of Discipline, Community Resources, School of Agricultural Sciences and Agribusiness, University of KwaZulu-Natal, Pietermaritzburg; and Director (Teaching and Learning) and Academic Director (Extension & Resource Management), Centre for Environment, Agriculture and Development, School of Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg.

¹The authors are aware that smallholders' involvement in the cultivation of cash crops is better represented by the income they earn from these crops. However, income was not assessed in this study.

²More than seven routes and railways from Nampula, Zambezia, Cabo Delgado and Niassa provinces, as well as from Malawi, converge in Cuamba. Most of the agribusiness companies are based in the city of Cuamba, giving this district a competitive advantage over the surrounding districts and allowing smallholders to employ as ganho-ganho people who come here in search of better conditions. Newcomers often first work as ganho-ganho before they cultivate their own fields in the following years.

³Lowland is land near a superficial or groundwater source that maintains enough water in the soil to sustain crop production during the dry season.

⁴With an average yield of 600 kg/ha, a household will need 1.2 ha per year to produce 715 kg of maize required for consumption (715 kg/year/600 kg/ha year) (Akesson, 1994).