Linkage between working conditions and wellbeing: Insight from migrant and native farmhands on Ghana's cocoa farms

Abstract

Working conditions and wellbeing (quality of life) could be linked, and they in turn enhance social and economic development. Nevertheless, working conditions of farmhands have largely been ignored in policy and research. We explored working conditions of migrant and native farmhands on Ghana's cocoa farms, and implications on wellbeing, using primary data from 600 respondents. Multidimensional Poverty Index, Department for International Development sustainable livelihood approach, World Food Programme asset score, Zellner's seemingly unrelated regression and multinomial logistic regression were adopted. Living standards, resilience, health and asset ownership of farmhands were generally low. Natives had higher living standards than migrants. However, migrants had better food security, and were more resilient to shocks than natives. Working and living conditions like years as a farmhand, closeness to social amenities, years migrant had stayed in community, type of migrant, being joined by a household member, working hours and days, type of agreement, category of farmhand, bonuses, satisfaction with working conditions, and income influence living standards, resilience, health and asset ownership. Thus, there is a link between working conditions and wellbeing of cocoa farmhands. Farmhands should be given long-term contracts, bonuses/incentives and personal protective equipment (PPE) by cocoa farmers. Government and private agencies should provide social amenities/infrastructure in cocoa-growing communities. Farmhands should do their own farms and join associations.

1. Introduction

Cocoa production contributes immensely to Ghana's economy [1,2]. However, it is mainly on small scale and highly labour-intensive [3,4]. Smallholder cocoa farmers, who form the majority, have low financial capacities to hire the required labour [5,6]. Therefore, most cocoa farmers use cheaper sources of labour like migrant, child and family labour [3,5,6]. Labour issues in cocoa production have been pertinent since colonial era and have since been linked to migratory labour [7,8]. Migration is a livelihood and survival strategy for many households [7,9]. Many people are motivated to migrate because they have the perception that working conditions at destination are poor better [7,10]. The major rationale for migration is lack of job opportunities and good wages at home [11]. Therefore, migration is linked to the quest for decent and well-paid employment opportunities [10].

Primarily, differences in economic opportunities, resource endowments, climatic conditions, and insufficiency or infertility of agricultural land have led to widespread labour migration from one region of Ghana to the other, especially from the northern part to the southern part [7]. Some of these migrants end up in cocoa growing localities to service as farm labourers [7]. Such migrants need good working conditions to enhance their safety/health and general wellbeing. Nevertheless, most migrants suffer bad working conditions such as poor remuneration, unsafe working environs, discrimination, no freedoms and rights [12]. Mostly, migrants experience long hours of work and lack of bargaining power [10]. Poor wages and other forms of bad conditions of work are widespread among migrants on cocoa plantations in Ghana [12]. There are instances where migrant farmworkers suffer food scarcity, accommodation, and other basic needs which hamper their living standards [13]. There could be employment discrimination as migrant workers could be the only labour employed in situations where the safety conditions are so poor [14]. Labour migration may present threats to migrants’ health since certain jobs expose them to workplace injuries and diseases [3,5,7].

To sustain the cocoa subsector, it is relevant to ensure satisfactory wellbeing of cocoa farmworkers. However, research, government and private sector interventions had not paid much attention to the wellbeing of informal labour in Ghana, predominantly, cocoa farmhands. Hence, studies on the working conditions of farmhands in Ghana's cocoa subsector are uncommon. Also, most migration studies focus on the economic implications of migration like remittances sent home [15,16]. To fill these gaps in literature, it is important to investigate the wellbeing of migrant and native farmhands on Ghana's cocoa farms, and the linkage between working conditions and wellbeing of farmhands. The specific objectives of the study are as follows. (1) To investigate and compare the wellbeing of migrant and native cocoa farmhands in Ghana in terms of living standards, resilience and health, food security, and asset ownership. (2) To examine the linkage between working conditions and wellbeing of cocoa farmhands (that is the implications of working conditions on the wellbeing of migrant and native farmhands).

2. Empirical studies on labour migration

Due to its relevance to household livelihood, agricultural and rural economy, and national development, there are many studies on labour migration: [[16], [17], [18], [19], [20], [21]]. There are environmental push and pull factors that determine migration flows, although these environmental factors do not usually act alone [21]. Most migrants relocate due to scarcity of fertile land, unreliable rainfall, climate change, low yield of crops and food insecurity in Northern Ghana [21]. Few farmers cited non-environmental reasons for migration like absence of non-farm income opportunities, family conflicts, witchcraft and desire to be independent [21]. Most employers prefer migrant to native labourers due to the willingness of the former to work for longer hours coupled with their greater obedience and non-complaining attitude compared with native workers [18]. The core reasons for labour migration are to access better job opportunities and good wages at destination [18]. According to the study, the main problems faced by migrants are: exploitation by intermediaries in the form of low wages; lack of permanent work; lack of bargaining power; and in-conducive working conditions.

[17] found that majority of migrants remit home. However, according to the study, migration has negative impact on food crop production in sending communities since remittances are not invested in agricultural production to offset the lost labour. It was found that a greater proportion of remittance is spent on food and school fees [22]. indicated that migration reduces both agricultural production and productivity in sending communities. According to Ref. [23] migration does not have any positive impact on farm production as it often serves as a means for agricultural households to stop farming.

[16,23,24] reported that remittances boost productive investments in sending communities. According to Ref. [24], remittances are used to cope with certain agricultural production shocks [25]. reported that migration and remittance have positive investment-promotion effects on agricultural production [20]. concluded that high remittances lead to investment in high-productivity crop and livestock farming, and preference for non-farm economic activities, increased consumption and leisure. The study found that migration and remittance alone are insufficient to transform subsistence agriculture into commercial agriculture since whenever remittance is adequately high to boost subsistence farming, there is a high likelihood that most households would neglect farming and engage in off season economic activities [23]. emphasized that outmigration has negative effect on traditional agriculture.

[26] asserted that migration has positive impact on diversification and agricultural investment since it minimizes credit and risk challenges faced by households; but the positive impact depends on amount of remittances sent [27]. found that migration increases adoption of high yielding crop varieties but this depends on farmers’ adoption behaviour [27]. revealed that impact of migration on agricultural investment depends on destination wage.

3. Research methodology

3.1. Sampling

This study was carried out in Western (now Western and Western North Regions) and Ashanti Regions of Ghana. This are the main cocoa-producing regions in Ghana. Primary data was obtained from migrant and native farmhands on cocoa farms, and cocoa farmers (200 each). Personal interviews were conducted using a semi-structured questionnaire.

Multistage sampling was used. In stage one, names of the main cocoa-growing districts in Western and Ashanti Regions were collected from Ghana Cocoa Board (COCOBOD). Purposive sampling was used in selecting two main cocoa districts from each of the two regions (due to financial constraint): Samreboi and Asawinso Cocoa Districts from Western Region, and Tepa and Bekwai Cocoa Districts from Ashanti Region. In stage two, names of main cocoa-growing villages in the sampled districts were taken from COCOBOD. In all, 200 villages were provided. Seven villages from each district were selected using simple random sampling procedure. In stage three, cocoa farmhands in each village were grouped into migrant and natives using stratified sampling. From each village, either seven or eight each of migrants, natives and cocoa farmers were selected, depending on availability. Sampled cocoa farmers who did not employ farmhands were ineligible for the interview, and were replaced. Eventually, 150 respondents, constituting 50 each of migrants, natives and cocoa farmers were selected from each district (300 from each region). Thus, the sample size was 600. During the data collection, informed consent was obtained from all participants. This study did not require any ethics committee clearance/approval since the work did not include human and animal specimen. However, the respondents were made to know the essence of the study and that they could opt out of the interview before or during the interview. Thus, participation was purely voluntary.

3.2. Assessing wellbeing of cocoa farmhands

Wellbeing indicators/livelihood outcomes used for the study are standard of living, resilience (low vulnerability), food security, health, and asset ownership with each having a number of parameters (Table 1, Table 2, Table 3). They were measured as scores. Multidimensional Poverty Index (MPI) approach for measuring non-monetary welfare developed by Ref. [28], sustainable livelihood method developed by Department for International Development (DFID) [29], and asset ownership developed by World Food Programme [30] were adapted for assessing farmhands’ wellbeing: standard of living [28], resilience [29], food security [28,29], health [28], and asset ownership [28,30].

Table 1.

Parameters and coding for living standard, resilience, and health scores.

Parameter	Coding (1 = high, 0.5 = moderate, 0 = deficient)	Score
Wellbeing indicator: standard of living
Improved drinking water	1 = farmhand has safe drinking water (pipe borne, borehole, sachet water) at home, or safe drinking water is less than 30 min-walk from home.
	0.5 = safe drinking water is more than 30 min-walk from home.
	0 = farmhand does not have access to safe drinking water.
Electricity	1 = farmhand has electricity at home.
	0.5 = there is electricity in labourer's community but no electricity at home.
	0 = no electricity in labourer's community.
Improved sanitation	1 = farmhand has improved toilet at home, and does not share with other household(s).
	0.5 = farmhand has toilet at home but shares with other household(s).
	0 = farmhand does not have toilet at home.
Flooring	1 = floor of farmhand's living room is made of tiles.
	0.5 = floor of farmhand's living room is made of cement.
	0 = floor of farmhand's living room is made of earth/mud.
Roofing material	1 = farmhand's house is roofed with tiles.
	0.5 = farmhand's house is roofed with iron sheet.
	0 = farmhand's house is roofed with grass/thatch.
Building material	1 = farmhand's house is built with cement block.
	0.5 = farmhand's house is built with mud brick.
	0 = farmhand's house is built with mud, mat or wood.
Cooking fuel	1 = farmhand's cooks with liquefied petroleum gas.
	0.5 = farmhand's cooks with charcoal.
	0 = farmhand's cooks with firewood.
Living standard score
Wellbeing indicator: resilience
Job security	1 = farmhand has permanent/long-term contract with principal.
	0 = farmhand has temporary/short-term contract with principal.
Credit accessibility in times of financial crisis	1 = yes; 0 = no
Farm ownership	1 = yes; 0 = no
Social capital	1 = farmhand belongs to more than one association.
	0.5 = farmhand belongs to one association.
	0 = farmhand does not belong to any association.
Resilience score
Wellbeing indicator: health
Ability to use health facility	1 = farmhand can afford medical/hospital bills when sick.
	0 = farmhand cannot afford medical/hospital bills when sick.
Child mortality	1 = no child of the farmhand had died five years before the survey.
	0 = a child of the farmhand had died five years before the survey.
Health score

Table 2.

Parameters and coding for food security score.

Parameter	Coding (1 = high, 0.5 = moderate, 0 = deficient)	Score
Cereals, roots and tubers	1 = farmhand had eaten a starchy staple in the past three days before the survey.
	0.5 = farmhand had eaten a starchy staple in the past seven days before the survey.
	0 = farmhand had not eaten a starchy staple in the past seven days before the survey.
Vegetables	1 = farmhand had eaten vegetables in the past three days before the survey.
	0.5 = farmhand had not eaten vegetables in the past seven days before the survey.
	0 = farmhand had not eaten vegetables in the past seven days before the survey.
Legumes	1 = farmhand had eaten legumes in the past three days before the survey.
	0.5 = farmhand had eaten legumes in the past seven days before the survey.
	0 = farmhand had not eaten legumes in the past seven days before the survey.
Fruits	1 = farmhand had eaten fruit in the past three days before the survey.
	0.5 = farmhand had not eaten fruit in the past seven days before the survey.
	0 = farmhand had not eaten fruit in the past seven days before the survey.
Meat or fish	1 = farmhand had eaten meat or fish in the past three days before the survey.
	0.5 = farmhand had eaten meat or fish in the past seven days before the survey.
	0 = farmhand had not eaten meat or fish in the past seven days before the survey.
Egg or milk	1 = farmhand had eaten egg or milk in the past three days before the survey.
	0.5 = farmhand had not eaten egg or milk in the past seven days before the survey.
	0 = farmhand had not eaten egg or milk in the past seven days before the survey.
Fat and oil	1 = farmhand had eaten fat or oil food in the past three days before the survey.
	0.5 = farmhand had not eaten fat or oil food in the past seven days before the survey.
	0 = farmhand had not eaten fat or oil food in the past seven days before the survey.
Supplementary foods (condiments)	1 = farmhand had eaten condiments (natural spices, biscuits, tea, alcoholic or non-alcoholic drinks) in the past three days before the survey.
	0.5 = farmhand had eaten condiments in the past seven days before the survey.
	0 = farmhand had not eaten condiments in the past three days before the survey.
Food availability	1 = farmhand have enough food all-year-round.
	0 = farmhand does not have enough food all-year-round.
Food security score

Table 3.

Asset score.

Asset	A	B	C
	Ownership	Functionality of asset	A × B
	1 = yes	1 = fully functioning
	0 = no	0.5 = partially functioning
		0 = not functioning, or farmhand does not own asset
Mobile phone
Radio
Television
Fan
Refrigerator
Sewing machine
Electric/gas stove
Means of transport (bicycle/motorbike/tricycle/car)
Asset score (sum of C)

Source: Adapted from [30].

This study focused mainly on livelihood outcomes in the DFID framework. The following proxies were used: income from being a cocoa farmhand for ‘more income,’ living standards for ‘increased wellbeing,’ resilience for reduced vulnerability,’ and dietary diversity and access and food availability all-year-round for ‘improved food security.’ Certain assets are required to accomplish good livelihood outcomes [29]. Thus, asset score [30] was adapted in measuring asset ownership by farmhands.

3.3. Assessing the linkage between working conditions and wellbeing of cocoa farmhands

The linkage between working conditions and wellbeing of cocoa farmhands (that is the implications of working conditions on wellbeing of migrant and native cocoa farmhands) were assessed with Zellner's seemingly unrelated regression (SURE) and multinomial logistic regression (MNL).

3.3.1. Zellner's SURE

Zellner's SURE estimator [31] simultaneously estimated the linkage between working conditions and the five wellbeing indicators of migrant and native farmhands: living standards, resilience, health, food security and asset ownership. Thus, there were five models in the SURE which were fitted using the same variables at the right-hand-side as shown in equations (1), (2), (3), (4), (5). Error terms of the five response variables could be correlated [32].

$$y_1=X_1{\beta }_1+{\epsilon }_1$$ (1)

$$y_2=X_2{\beta }_2+{\epsilon }_2$$ (2)

$$y_3=X_3{\beta }_3+{\epsilon }_3$$ (3)

$$y_4=X_4{\beta }_4+{\epsilon }_4$$ (4)

$$y_5=X_5{\beta }_5+{\epsilon }_5$$ (5)

where; $y_i$ represents the response variables ($y_1$ is living standards, $y_2$ is resilience, $y_3$ is health, $y_4$ is food security, and $y_5$ is asset ownership); $X_i$ represents explanatory variables (Table 1); $\beta$ represents parameters to be estimated; $\epsilon ={\left[{\epsilon }_1^{\prime },{\epsilon }_2^{\prime },{\epsilon }_3^{\prime },{\epsilon }_4^{\prime },{\epsilon }_5^{\prime }\right]}^{\prime }$; $E\left[\epsilon |X_1,X_2,X_3,X_4,X_5\right]=0\text{;}$ $E\left[\epsilon {\epsilon }^{\prime }|X_1,X_2,X_3,X_4,X_5\right]=\Omega \text{.}$ Indices were computed separately for each wellbeing indicator (see Table 2, Table 3, Table 4). These were used as response variables in the SURE.

Table 4.

Variables for Zellner's SURE and MNL.

Variable	Measurement
Response variable
SURE: (1) Living standard (2) Resilience (3) Food security (4) Health (5) Asset ownership	Continuous: Computed scores
MNL: Level of: (1) living standard (2) resilience (3) food security (4) health (5) asset ownership	Categorical: 0 = low/deficient, 1 = moderate, 2 = high
Explanatory variable
Sex	1 = male, 0 = female
Age	Years
Education	Years
Dependents	Number
Off-farm	1 = yes, 0 = no
Years as a cocoa farmhand	Years
Closeness to social amenities	Kilometres
Years migrant had stayed in community	Years
Type of migrant	1 = permanent, 0 = temporary/seasonal
Household member join migrant	1 = yes, 0 = no
Working days	Number of days farmhand works a week
Working hours	Number of hours farmhand works a day
Nature of contract	1 = permanent, 0 = temporary
Category of farmhand	1 = caretaker, 0 = otherwise
Farmers a farmhand serve	Number
Bonuses	1 = principal gives farmhand bonuses, 0 = otherwise
Satisfaction with conditions of work	Mean score for satisfaction with conditions of work
Income from being a cocoa farmhand	US$a

^aGhana cedis (GH₵) is Ghana's currency; US$1 = GH₵5.74.

3.3.2. Multinomial logistic regression (MNL)

Scores were computed using the number of parameters for each wellbeing indicator. The score for each wellbeing indicator was further divided into three to obtain high, moderate and low living standard, resilience, health, food security and asset scores. Potential living standard score is seven, since it has seven parameters (Table 1). Thus, 0.00–2.33 = low/deficient, 2.34–4.67 = moderate, and 4.68–7.00 = high. Potential resilience score is four, since it has four parameters (Table 1). Thus, 0.00–1.33 = low/deficient, 1.34–2.67 = moderate, 2.68–4.00 = high. Potential health score is two, since it has two parameters (Table 1). Thus, 0.00–0.67 = low/deficient, 0.68–1.33 = moderate, 1.34–2.00 = high. Potential food security score is nine, since it has nine parameters (Table 2). Thus, 0.00–3.00 = low/deficient, 3.01–6.00 = moderate, 6.01–9.00 = high. Potential asset score is eight, since eight assets were used (Table 3). Thus, 0.00–2.67 = low/deficient, 2.68–5.34 = moderate, 5.35–8.00 = high. Likewise, potential overall wellbeing score (based on the five indicators) is 30. Thus, 0–10 = low/deficient, 10.01–20 = moderate, 20.01–30 = high. These levels of wellbeing were used as response variables for the MNL. Therefore, there were three outcomes, 0, 1, 2, recorded in each $y\text{.}$ MNL estimates a set of coefficients, ${\beta }^{\left(0\right)},{\beta }^{\left(1\right)},and{\beta }^{\left(2\right)}\text{,}$ corresponding to each outcome as shown in equations (6), (7), (8) [32]:

$$\mathrm{Pr}\left(y=0\right)=\frac{e^{X{\beta }^{\left(0\right)}}}{e^{X{\beta }^{\left(0\right)}}+e^{X{\beta }^{\left(1\right)}}+e^{X{\beta }^{\left(2\right)}}}$$ (6)

$$\mathrm{Pr}\left(y=1\right)=\frac{e^{X{\beta }^{\left(1\right)}}}{e^{X{\beta }^{\left(0\right)}}+e^{X{\beta }^{\left(1\right)}}+e^{X{\beta }^{\left(2\right)}}}$$ (7)

$$\mathrm{Pr}\left(y=2\right)=\frac{e^{X{\beta }^{\left(2\right)}}}{e^{X{\beta }^{\left(0\right)}}+e^{X{\beta }^{\left(1\right)}}+e^{X{\beta }^{\left(2\right)}}}$$ (8)

The logistic function is expressed in equation (9):

$$f\left(x\right)=\frac{L}{1+e^{-k\left(x-x_0\right)}}$$ (9)

where; $x_0$ denotes the $x$ value of the sigmoid's midpoint; $L$ denotes supremum of the values of the function; and $k$ denotes the logistic growth rate/steepness of the curve. The base categories differed for the five wellbeing indicators, since not all indicators had the three wellbeing levels. The base outcome for living standard was low. No respondents had high level of living standard. The base outcome for resilience was low. Only 0.25% had low food security. Thus, high was used as base outcome, and compared with moderate. For health, only 1.75% were low/deficient. Thus, high was used as base outcome. The base outcome for asset ownership was low. Table 4 presents variables used for MNL.

4. Results and discussion

4.1. Background information of farmhands

Table 5 presents background information of migrant and native farmhands on Ghana's cocoa plantations. The test statistics shows significant differences in characteristics of migrants and natives. Males were dominant in labour provision and cocoa farming, which agrees with the findings of [5,7]. Farmhands were in their late 30s and farmers were in their 50s. Male dominance and youthful age of farmhands might be as a result of the difficult nature of cocoa production [33]. Natives (seven) recorded higher years of schooling than migrants (five). In agreement with this [1,4,33], reported that cocoa farmers in Ghana have lower levels of formal education. They recorded the same household sizes (four each). Farmhands had 10 years of experience while farmers had 20 years of experience. Thus, they have amassed experience in the industry. There were more natives (46%) with off-farm occupations than migrants (39%). Migrants had been in receiving communities for about 16 years. Migrants were either temporary/seasonal (40%) or permanent (60%) in receiving communities. Half of migrants had household member(s) joined them in receiving communities. These findings generally agree with those of [3,5].

Table 5.

Background information of migrant and native cocoa farmhands.

Variable	Description	Migrants (n = 200)	Non-migrants (n = 200)	Test statistics	Aggregate (N = 400)	Cocoa farmers (n = 200)
Sex	Males	84.50	75.00	8.77***	79.75	76.00
	Females	15.50	25.00		20.25	24.00
Age	Mean	37.09	36	93.71***	36.55	51.50
	Standard deviation	7.84	7.42		7.64	12.72
	Minimum	18	21		18	20
	Maximum	59	52		59	88
Education	Mean	5.49	6.63	24.56***	6.06	8.39
	Standard deviation	4.49	4.56		4.55	5.97
	Minimum	0	0		0	0
	Maximum	14	18		18	18
Household size	Mean	4.46	4.78	36.70***	5.12	5.77
	Standard deviation	2.72	1.97		2.40	2.24
	Minimum	1	1		1	1
	Maximum	15	10		15	15
Dependents	Mean	4.60	3.71	26.13***	4.16	4.69
	Standard deviation	2.87	2.38		2.67	2.38
	Minimum	0	0		0	0
	Maximum	13	13		13	14
Years on cocoa farm as a farmhand	Mean	10.93	10.03	29.66***	10.48	20.31
	Standard deviation	6.97	6.36		6.68	13.21
	Minimum	0.08	1		0.08	1
	Maximum	35	30		35	65
Off-farm occupation	Yes	38.50	46.00	2.29**	42.25	46.50
	No	61.50	54.00		57.75	53.50
Years in community as a migrant	Mean	16.27
	Standard deviation	12.44
	Minimum	0.08
	Maximum	57
Type of migrant	Permanent	60.50
	Temporary/seasonal	39.50
Whether household member had joined migrant	Yes	51.00
	No	49.00
Membership of faith-based organizations	Yes	85.50	71.50	9.64***	78.50	85.50
	No	14.50	28.50		21.50	14.50
Membership of tribal associations	Yes	14.00	9.50	13.28***	11.75	14.50
	No	86.00	90.50		88.25	85.50
Membership of fun clubs	Yes	2.50	1.50	18.64***	2.00	1.00
	No	97.50	98.50		98.00	99.00
Membership of farmer/labourers' associations	Yes	19.50	10.00	12.29***	14.75	41.00
	No	80.50	90.00		85.25	59.00
Social capital	Mean	1.27	1.00	12.92***	1.13	1.44
	Standard deviation	0.77	0.71		0.75	0.77
	Minimum	0	0		0	0
	Maximum	4	4		4	4

***and ** denote significance at 1% and 5% respectively.

Membership of faith-based organizations, tribal associations, fun clubs and farmer/labourers' associations were proxy for social capital (Table 5). There were more migrants (86%) who belonged to faith-based organizations like churches and Islam than natives (72%). Few migrants (14%) and natives (10%) belonged to tribal associations, and less than 5% belonged to fun clubs. More migrants (20%) have joined farmer/labourers’ associations than natives (10%). Averagely, farmhands were members of one out of the four groups. However, some did not belong to any association, while other belonged to all. There is low level of association membership among cocoa farmers in Ghana [3,5,33].

4.2. Wellbeing of cocoa farmhands

Table 6, Table 7, Table 8, Table 9 show wellbeing of migrant and native cocoa farmhands. Table 6 shows living standards, resilience and health scores. The test statistics show significant differences in these wellbeing parameters between migrants and natives. Natives (71%) had access to safe drinking water than migrants (61%). Such farmhands had pipe or borehole at home, or these water sources were shorter than 30 min-walk from home. Most farmhands did not have these water sources at home but fetched/bought at locations shorter than 30 min-walk from home. Less than 5% of farmhands were drinking safe water but had to walk more than 30 min from home to fetch/buy. The deficiency level of migrants (37%) and natives (26%) in terms of access to safe drinking water implies that some farmhands were drinking from wells, springs, dams and rivers/streams. These unsafe water sources are harmful to their health. Normally, migrants in cocoa growing communities occupy ‘inferior’ areas compared with natives which reduces their access to community resources like potable water [34,35].

Table 6.

Living standards, resilience and health scores.

Parameter	Level of living standard, resilience, health	Proportion of farmhands (%)
		Migrants	Natives	Test statistics	Aggregate
Living standard
Improved drinking water	High	60.50	70.50	−5.39***	65.50
	Moderate	3.00	4.00		3.50
	Deficient	36.50	25.50		31.00
Electricity	High	41.00	48.50	0.04	44.75
	Moderate	8.00	12.50		10.25
	Deficient	51.00	39.00		45.00
Improved sanitation	High	5.00	7.50	6.44***	6.25
	Moderate	51.50	49.00		50.25
	Deficient	43.50	43.50		43.50
Flooring	High	0.00	0.00	1.89*	0.00
	Moderate	89.00	91.50		90.25
	Deficient	11.00	8.50		9.75
Roofing material	High	0.00	0.00	0.88	0.00
	Moderate	95.50	95.50		95.50
	Deficient	4.50	4.50		4.50
Building material	High	11.50	15.50	6.42***	13.50
	Moderate	30.50	42.00		36.25
	Deficient	58.00	42.50		50.25
Cooking fuel	High	0.00	0.00	18.12***	0.00
	Moderate	5.50	11.50		8.50
	Deficient	94.50	88.50		91.50
Living standard score	High	0.00	0.00		0.00
	Moderate	40.00	50.50		45.25
	Deficient	60.00	49.50		54.75
	Mean	1.98	2.23	32.04***	2.10
	Standard deviation	0.88	0.98		0.93
	Minimum	0.50	0.00		0.00
	Maximum	4.00	4.50		4.50
Resilience/low vulnerability
Job security	High	39.00	43.50	2.55**	41.25
	Deficient	61.00	56.50		58.75
Credit accessibility	High	58.00	46.50	−0.60	52.25
	Deficient	42.00	53.50		47.75
Farm ownership	High	81.00	79.50	−9.37***	80.25
	Deficient	19.00	20.50		19.75
Social capital	High	31.50	20.50	−1.32	26.00
	Moderate	56.50	56.50		56.50
	Deficient	12.00	23.00		17.50
Resilience score	High	37.50	24.50		31.00
	Moderate	50.50	60.50		55.50
	Deficient	12.00	15.00		13.50
	Mean	2.38	2.18	−31.67***	2.28
	Standard deviation	0.98	0.93		0.96
	Minimum	0.5	0.00		0.00
	Maximum	4.00	4.00		4.00
Health
Ability to use health facility	High	79.50	81.50	−9.67***	80.50
	Deficient	20.50	18.50		19.50
Child mortality	High	93.50	92.00	−14.98***	92.75
	Deficient	6.50	8.00		7.25
Health score	High	86.50	86.75		86.63
	Deficient	13.50	13.25		13.37
	Mean	1.73	1.74	−35.59***	1.73
	Standard deviation	0.47	0.50		0.48
	Minimum	0.00	0.00		0.00
	Maximum	2.00	2.00		2.00

***, ** and * denote significance at 1%, 5% and 10% respectively.

Table 7.

Food security.

Parameter	Level of food security	Proportion of labourers (%)
		Migrants	Natives	Test statistics	Aggregate
Cereals, roots and tubers	High	100.00	100.00	−19.98***	100.00
	Moderate	0.00	0.00		0.00
	Deficient	0.00	0.00		0.00
Vegetables	High	100.00	100.00	−19.98***	100.00
	Moderate	0.00	0.00		0.00
	Deficient	0.00	0.00		0.00
Legumes	High	54.50	46.50	−4.25***	50.50
	Moderate	29.00	26.50		27.75
	Deficient	16.50	27.00		21.75
Fruits	High	49.00	52.00	−4.18***	50.50
	Moderate	29.00	23.00		26.00
	Deficient	22.00	25.00		23.50
Meat or fish	High	98.50	99.50	−19.70***	99.00
	Moderate	1.00	0.50		0.75
	Deficient	0.50	0.00		0.25
Fat and oil	High	87.00	94.00	−17.53***	90.50
	Moderate	10.00	6.00		8.00
	Deficient	3.00	0.00		1.50
Egg or milk	High	2.00	2.50	18.08***	2.25
	Moderate	1.00	1.50		1.25
	Deficient	97.00	96.00		96.50
Supplementary foods/condiments	High	56.00	61.00	−3.53**	58.50
	Moderate	8.00	5.50		6.75
	Deficient	36.00	33.50		34.75
Food availability all-year-round	High	23.50	10.00	9.83***	16.75
	Deficient	76.50	90.00		83.25
Food security score	High	41.50	36.00		38.75
	Moderate	58.00	64.00		61.00
	Deficient	0.50	0.00		0.25
	Mean	6.10	5.97	−99.13**	6.03
	Standard deviation	0.99	0.93		0.97
	Minimum	3.00	4.00		3.00
	Maximum	8.00	9.00		9.00

*** and ** denote significance at 1% and 5% respectively.

Table 8.

Asset ownership/score.

Asset	Labourer	Ownership (%)	Test statistics	Percentage (%) functionality of asset
				Fully functioning	Partially functioning	Not functioning or does not own asset	ANOVA (F-statistics)
Mobile phone	Migrants	82.00		76.00	5.00	19.00
	Natives	79.50		79.00	0.00	21.00
	Aggregate	80.75	−9.50***	77.50	2.50	20.00	5.25***
Radio	Migrants	81.00		74.00	6.50	19.50
	Natives	81.50		81.00	0.50	18.50
	Aggregate	81.25	−9.87***	77.50	3.50	19.00	5.60***
Television	Migrants	27.50		22.50	3.00	74.50
	Natives	44.50		44.00	0.50	55.50
	Aggregate	36.00	4.45***	33.25	1.75	65.00	12.13***
Fan	Migrants	14.50		13.00	1.00	86.00
	Natives	31.00		29.50	1.50	69.00
	Aggregate	22.75	9.29***	21.25	1.25	77.50	8.67***
Refrigerator	Migrants	3.00		2.50	0.50	97.00
	Natives	5.00		4.50	0.00	95.50
	Aggregate	4.00	17.41***	3.50	0.25	96.25	1.08
Sewing machine	Migrants	3.00		2.00	1.00	97.00
	Natives	6.50		3.00	2.00	95.00
	Aggregate	4.75	17.15***	2.50	1.50	96.00	0.55
Electric/gas stove	Migrants	0.50		0.00	0.50	99.50
	Natives	3.50		3.50	0.00	96.50
	Aggregate	2.00	19.00***	1.75	0.25	98.00	4.10**
Means of transport	Migrants	35.50		33.50	2.00	64.50
	Natives	39.50		35.50	4.50	60.00
	Aggregate	37.50	3.66***	34.50	3.25	62.25	1.18
Asset score	Description	Migrants	Natives	Aggregate	Test-statistics
	Mean	2.33	2.84	2.59	−29.29***
	Standard deviation	1.25	1.55	1.43
	Minimum	0.00	0.00	0.00
	Maximum	6.00	7.00	7.00
	High (%)	2.00	4.00	3.00
	Medium (%)	37.00	49.00	43.00
	Deficient/low (%)	61.00	47.00	54.00

*** and ** denote significance at 1% and 5% respectively.

Table 9.

Overall wellbeing score for farmhands.

Description	Migrants	Natives	Aggregate	Test-statistics
Mean	14.51	14.96	14.73	−97.45a
Standard deviation	2.70	3.11	2.92
Minimum	7.50	6.50	6.50
Maximum	21.50	24.00	24.00
High (%)	0.50	4.00	2.25
Medium (%)	91.50	91.00	91.25
Deficient/low (%)	8.00	5.00	6.50

^adenotes significance at 1%.

Less than half of migrants (41%) and natives (49%) had high access to electricity. Such farmhands had electricity at home (Table 6). For one-tenth of farmhands, there were electricity in their communities but they did not have electricity at home. These were farmhands who could not afford the initial cost of bringing electricity to their houses and/or pay the monthly electricity bills. Thus, some cocoa farmers pay monthly electricity bills for their farmhands which is a private benefit to farmhands [7]. However, 51% of migrants and 39% of natives were deprived of electricity. Such farmhands did not have electricity at home and there were no electricity in their communities. Thus, their sources of lighting were kerosene lamp, torch/flashlight and candle. Such farmhands were unable to use electrical gadgets/devices, though they mainly relies on nearby communities to recharge their mobile phones. The foregoing implies that migrants were deprived of safe drinking water and electricity than natives.

Less than one-tenth of farmhands had high access to improved sanitation. These were farmhands with improved toilet (pan/covered latrine or Kumasi Ventilated Improved Pit (KVIP) latrine) at home, and did not share with other household(s). No farmhand was using water closet toilet. Half of farmhands had toilet at home but shared with other household(s). Also, 44% each of migrants and natives did not have toilet at home; and were therefore, deprived. Such farmhands use toilet of other households or resort to open defecation. This implies that deficiency level in improved sanitation (toilet) was high.

No farmhand had tiles as flooring material in living room (Table 6). About 90% each of migrants and natives was moderate in flooring materials. The floors of their living rooms were made of cement. Less than 10% had their living rooms made of earth/mud; and were therefore, deprived. Furthermore, no farmhand had his/her house roofed with tiles. About 96% each of migrants and natives had their houses roofed with iron sheets/aluminium; and were therefore, moderate in roofing materials. About 5% each of migrants and natives were deprived in roofing materials. Such farmhands had their houses roofed with grass/thatch. Less than 15% of farmhands had their houses built with cement block. About 31% of migrants and 42% of natives had their houses built with mud brick; and were therefore, moderate in building materials. More migrants (58%) deprived in building materials than natives (43%). These were farmhands whose houses were built with mud. No farmhand was cooking with liquefied petroleum gas (LPG). Less than 10% were rated moderate in cooking fuel. These farmhands were cooking with charcoal. Migrants (95%) were deprived of cooking fuel than native (89%). Such farmhands were cooking with firewood. To reduce deforestation through tree cutting for charcoal and firewood, LPG is highly recommended. However, low income deter farmhands from using LPG.

The overall living standard score shows that no farmhand scored high in all the seven living standard parameters (Table 6). More natives (51%) had moderate living standards than migrants (40%). However, more migrants (60%) had deficient living standards than natives (50%). The mean shows that migrants had high in 1.98 out of the seven living standard parameters, while natives had 2.23. The minimum-maximum continuum reveals that some farmhands were not high in any living standard parameters while the highest was five. The foregoing implies that cocoa farmhands suffer low living standards, and natives enjoy better living standards than migrants. Low income/poor people work as farmhands, and farmhands receive lower remunerations compared with other workers. Most migrants are temporary/seasonal with the intention of returning home which makes it difficult for them to obtain good residence [34,35]. Migrants experience poor living conditions, including poor housing, than natives [12]. Thus, most temporary migrants on cocoa farms live moderate lives by building/residing in temporary structures like houses built and floored with mud, and roofed with thatch which are less expensive. Migrants live in cocoa farm farmsteads than indigenes. This reduces migrants’ access to electricity and potable water.

The second wellbeing indicator is farmhands' resilience to shocks or unexpected outcomes like financial crises, low cocoa yield or sickness. We used four parameters to assess farmhands’ resilience or vulnerability to shocks: job security, credit accessibility, farm ownership and social capital (Table 6). Natives (44%) had higher job security than migrants (39%). These were farmhands with permanent/long-term contracts. Permanent/long-term contracts assure farmhands of reliable income source. Such farmhands could plan with their income better than those with temporary/short-term contracts. Migrants (58%) had higher access to credit than natives (47%). Farmhands with access to credit could borrow during crises. This makes them less vulnerable to shocks. Most farmhands (80%) had their own farms. Such farmhands are more resilient to shocks since they rely on proceeds from their farms during income or food crises.

Social capital refers to membership of faith-based organizations (Christianity or Islam), tribal associations, fun clubs and farmer/labourers’ associations. Migrants (32%) had higher social capital than natives (21%). Equal share (57%) of migrants and natives belonged to one association. Few farmhands did not belong to any association. Farmhands with social networks assist each other during crises in the forms of resource, information and knowledge sharing [[36], [37], [38]]. reported that social capital enables people to cope better with food challenges. Though intangible, social networks help people in crises to get psychological supports [39]. These highlight the relevance of social capital in building resilience to shocks. The overall resilience score reveals that migrants (38%) had higher resilience than natives (25%). The mean indicates that farmhands had high resilience in two out of the four parameters. Thus, farmhands had moderate resilience to shocks/challenges.

The third wellbeing indicator is health (Table 6). Health was assessed with two parameters: health facility usage and child mortality. Most farmhands (80%) had the ability to use health facilities. These were farmhands who could afford medical bills when a household member was sick. However, some farmhands had challenges with lack of health centres in their communities. This was worst with farmhands residing in cocoa farmsteads. Most farmhands had subscribed to the National Health Insurance Scheme to aid affordability of medical bills. For child mortality, 7% had a child died five years before the study. The overall health score reveals that 87% of farmhands had high health in terms of health facility usage and child mortality.

The fourth wellbeing indicator is food security (Table 7). Food security was assessed with dietary diversity and access and food availability all-year-round. Dietary diversity and access was explored with previous three and seven days before the study as reference periods since these are appropriate timeframes for determining habitual diet [40,41]. The test statistics show significant differences in food security parameters between migrants and natives. All the farmhands had eaten starchy staples (cereals, roots or tubers) and vegetables in the past three days before the survey. Thus, none was moderate or deficient. Starchy staples constitute the most prominent food in most African countries [42]. Also, since farmhands are largely low income/poor, they consume more starchy staples (Bennett's law1). Vegetables are eaten regularly because of their prominence in almost every diet. In Ghana, vegetables are used as ingredients for preparing soup or stew that accompanies almost every food [43]. This results agree with [44] who reported that vegetables and starchy foods constitute greater than half of food costs of Ghanaian households.

Most farmhands had not eaten any legume or fruit in the past three (moderate) or seven (deficient) days before the survey, signifying high deficiency level (Table 7). Due to low income, fruit consumption is low in developing countries [45,46]. Few farmhands were deficient in the consumption of fish and meat (less than 1%) and fat and oil foods (less than 2%). Thus, most farmhands had eaten these foods in the past three or seven days before the survey. More than 95% of farmhands were deficient in the consumption of egg and milk. Such farmhands had not eaten egg or milk in the previous three or seven days before the study. This suggests high deficiency in the consumption of other protein sources apart from fish and meat. Greater than half of migrants (56%) and natives (61%) had eaten supplementary foods (tea, alcoholic or non-alcoholic beverages) in the past three days before the survey.

Natives (90%) experienced food inadequacy than migrants (77%); and were thus, deficient (Table 7). Farmhands with enough food all-year-round could be those with their own farms [13,14,47]. asserted that majority of migrants are likely to experience food scarcity within certain periods of the year. The overall food security score reveals that more than half of farmhands were moderate in food security. The mean indicates that farmhands scored high in six of the nine food security parameters.

The fifth wellbeing indicator is asset ownership (Table 8). The test statistics show significant differences in asset ownership between migrants and natives. Most farmhands (80%) had mobile phones and radios of which majority were fully functioning. Few farmhands owned television (36%), fan (23%) and means of transport (38%) like bicycle, motorbike, tricycle or car. Refrigerators, sewing machines and electric/gas stoves were owned by less than 5% of farmhands. The mean asset score for migrants (2.33) and natives (2.84) were very low, compared with the potential asset score of eight. This implies that farmhands owned three out of the eight assets. The minimum asset score implies that some farmhands did not own any of the assets or all assets owned were not at functioning. The mean asset score is lower than [30] in a baseline survey conducted in Ghana (5.0). Few farmhands (5%) were ranked high in asset ownership. Such farmhands owned more than five fully functioning assets. Nevertheless, more than half were deficient in asset ownership, since they owned less than three fully functioning assets. This suggests low asset ownership among cocoa farm labourers.

Table 9 shows overall wellbeing. The test statistic shows significant differences between the wellbeing of migrants and natives. The potential overall wellbeing score is 30. The average overall wellbeing score of migrants and natives was the same (15). This implies that farmhands scored high in half of the wellbeing indicators/parameters. The minimum-maximum continuum shows that some farmhands had high in only six of the 30 wellbeing parameters, while others had high in 24 parameters. Less than 3% were ranked high in wellbeing. Such farmhands had high in more than 20 parameters. Also, more than 90% had medium/moderate wellbeing. Such farmhands had high in 10–20 parameters. However, 7% were deficient in wellbeing, since they had high in less than 10 parameters. Therefore, generally, cocoa farmhands had moderate wellbeing.

4.3. Linkage between working conditions and wellbeing of cocoa farmhands

Table 10a, Table 10b, Table 11 shows Zellner's SURE estimator and MNL, respectively, for the linkage between working conditions and wellbeing of cocoa farmhands (implications of working conditions on migrant and farmhands' wellbeing). R-squared and pseudo R-squared in SURE and MNL, respectively, suggest that working conditions contribute to farmhands' wellbeing. The chi-squared and likelihood ratio chi-squared are significant, implying that working conditions jointly contribute to wellbeing.

Table 10a.

Zellner's SURE for the linkage between working conditions and wellbeing of cocoa farmhands.

Explanatory variable	Coefficient (z-statistics)
	Living standards			Resilience			Food security
	Migrants	Natives	Aggregate	Migrants	Natives	Aggregate	Migrants	Natives	Aggregate
Sex	−0.771 (−2.09)**	0.806 (1.41)	−0.214 (−0.69)	−0.144 (−0.48)	0.444 (1.05)	0.043 (0.18)	−0.683 (−1.61)	−0.199 (−0.33)	−0.473 (−1.40)
Age	0.009 (0.70)	−0.003 (−0.19)	0.0002 (−0.02)	−0.012 (−1.10)	0.008 (0.67)	−0.003 (−0.39)	−0.030 (−1.97)**	−0.011 (−0.67)	−0.018 (−1.61)
Education	−0.004 (−0.26)	0.022 (1.31)	0.019 (1.59)	−0.008 (−0.64)	0.010 (0.76)	−0.001 (−0.07)	0.029 (1.58)	−0.008 (−0.47)	0.012 (0.93)
Dependents	−0.060 (−1.97)**	0.026 (0.75)	−0.022 (−0.97)	0.013 (0.53)	0.057 (2.25)**	0.043 (2.45)**	−0.015 (−0.44)	0.038 (1.05)	0.014 (0.59)
Off-farm	0.053 (0.37)	0.180 (1.11)	0.121 (1.11)	0.104 (0.89)	−0.039 (−0.32)	0.071 (0.83)	0.182 (1.10)	−0.109 (−0.63)	0.037 (0.31)
Years as a farmhand	0.018 (1.25)	−0.014 (−0.83)	0.008 (0.77)	0.037 (3.06)***	0.0002 (0.01)	0.023 (2.71)***	−0.003 (−0.20)	−0.004 (−0.22)	−0.001 (−0.09)
Closeness to social amenities	−0.007 (−2.62)***	−0.008 (−4.32)***	−0.008 (−5.55)***	0.001 (0.06)	0.001 (0.91)	0.0001 (−0.01)	0.0004 (0.16)	0.001 (0.66)	0.001 (0.94)
Years in community as a migrant	−0.002 (−0.27)			0.002 (0.36)			−0.001 (−0.16)
Type of migrant	0.047 (0.28)			−0.006 (−0.05)			0.303 (1.58)
Household member join migrant	0.156 (1.06)			0.165 (1.37)			0.381 (2.24)**
Working days	0.324 (2.67)***	0.298 (2.82)***	0.322 (3.94)***	0.002 (0.02)	−0.102 (−1.30)	−0.034 (−0.53)	0.011 (0.08)	0.376 (3.36)***	0.233 (2.63)***
Working hours	0.110 (2.16)**	0.072 (1.13)	0.093 (2.35)**	0.077 (1.86)*	0.081 (1.72)*	−0.013 (−0.42)	0.116 (1.97)**	0.171 (2.53)**	0.138 (3.22)***
Type of contract	0.086 (0.56)	0.093 (0.53)	0.160 (1.41)	1.052 (8.48)***	0.755 (5.80)***	0.864 (9.74)***	−0.130 (−0.74)	0.373 (2.01)**	0.190 (1.54)
Category of farmhand	0.495 (2.75)***	0.243 (1.42)	0.327 (2.87)***	−0.266 (−1.82)*	0.006 (0.05)	−0.004 (−0.04)	0.205 (0.99)	0.269 (1.48)	0.203 (1.65)*
Number of farmers a farmhand serves	−0.006 (−0.56)	−0.007 (−0.53)	−0.012 (−1.37)	0.003 (0.31)	−0.011 (−1.09)	0.0008 (0.01)	0.008 (0.67)	−0.004 (−0.29)	0.002 (0.17)
Bonuses	−0.151 (−1.08)	0.082 (0.53)	−0.038 (−0.36)	0.210 (1.85)*	0.062 (0.54)	0.148 (1.81)*	−0.144 (−0.90)	−0.253 (−1.57)	−0.143 (−1.26)
Satisfaction with conditions of work	1.494 (0.21)	−8.662 (−0.89)	−4.488 (−0.77)	2.153 (0.37)	−7.051 (−0.97)	−1.637 (−0.36)	11.043 (1.33)	0.799 (0.08)	7.019 (1.12)
Income from being a cocoa farmhand	0.000 (1.69)*	0.0002 (0.80)	0.0002 (1.05)	0.002 (0.88)	0.0003 (−1.41)	0.0002 (0.21)	0.0005 (1.79)*	0.0003 (1.12)	0.0004 (1.81)*
Constant	3.268 (3.74)***	2.618 (2.59)**	3.393 (5.08)***	1.531 (2.16)**	2.234 (2.99)***	1.890 (3.63)***	6.222 (6.19)***	7.245 (6.77)***	7.121 (9.85)***
Parms	18	15	15	18	15	15	18	15	15
R-squared	0.315	0.386	0.277	0.515	0.349	0.366	0.184	0.264	0.153
Chi-squared	57.42***	73.61***	93.75***	132.48***	62.68***	141.40***	28.17*	41.94***	44.15***

***, ** and * denote significance at 1%, 5% and 10% respectively.

Table 10b.

Zellner's SURE for the linkage between working conditions and wellbeing of cocoa farmhands.

Explanatory variable	Coefficient (z-statistics)
	Health			Asset ownership
	Migrants	Natives	Aggregate	Migrants	Natives	Aggregate
Sex	−0.186 (−0.90)	−0.367 (−1.18)	−0.147 (−0.87)	1.131 (2.33)**	0.456 (0.48)	1.161 (2.40)**
Age	−0.003 (−0.36)	0.003 (0.42)	−0.002 (−0.29)	−0.014 (−0.82)	0.029 (1.14)	−0.001 (−0.03)
Education	0.004 (0.43)	0.012 (1.31)	0.009 (1.40)	0.003 (0.15)	0.027 (0.97)	0.036 (2.01)*
Dependents	−0.022 (−1.27)	−0.055 (−2.93)***	−0.037 (−2.99)***	0.060 (1.51)	0.060 (1.04)	0.044 (1.28)
Off-farm	0.010 (0.12)	0.208 (2.36)**	0.108 (1.81)*	0.497 (2.61)***	−0.026 (−0.10)	0.324 (1.91)*
Years as a farmhand	−0.001 (−0.11)	−0.001 (−0.17)	0.001 (0.18)	0.034 (1.73)*	−0.035 (−1.30)	0.004 (0.21)
Closeness to social amenities	−0.003 (−2.17)**	0.0008 (−0.09)	−0.001 (−1.06)	−0.006 (−1.65)*	−0.010 (−3.64)***	−0.008 (−3.99)***
Years in community as a migrant	0.007 (1.53)			0.003 (0.26)
Type of migrant	0.156 (1.67)*			0.385 (1.75)*
Household member join migrant	0.054 (0.66)			0.316 (1.62)
Working days	0.021 (0.32)	0.097 (1.70)*	−0.044 (−1.00)	−0.128 (−0.80)	0.644 (3.69)***	0.406 (3.22)***
Working hours	−0.022 (−0.77)	−0.017 (−0.48)	0.001 (0.05)	0.062 (0.91)	0.013 (0.12)	0.058 (0.95)
Type of contract	−0.062 (−0.72)	−0.002 (−0.02)	−0.030 (−0.48)	0.255 (1.26)	0.042 (0.14)	0.323 (1.84)*
Category of farmhand	0.185 (1.83)*	0.097 (1.05)	0.136 (2.19)**	0.289 (1.22)	0.497 (1.76)*	0.101 (0.57)
Number of farmers a farmhand serves	0.007 (1.09)	0.001 (0.16)	0.001 (0.24)	−0.002 (−0.15)	−0.025 (−1.17)	−0.021 (−1.58)
Bonuses	0.112 (1.43)	0.166 (2.01)**	0.130 (2.30)**	−0.105 (−0.57)	0.061 (0.24)	−0.091 (−0.56)
Satisfaction with conditions of work	−5.588 (−1.38)	12.422 (2.34)**	0.820 (0.26)	0.691 (0.07)	8.942 (0.55)	−2.871 (−0.32)
Income from being a cocoa farmhand	0.0001 (0.85)	0.0007 (−0.50)	0.0004 (0.50)	0.0001 (3.36)***	0.0002 (0.56)	0.0003 (1.16)
Constant	1.772 (3.62)***	2.356 (4.31)***	1.896 (5.24)***	0.751 (0.65)	5.487 (3.29)***	3.189 (3.09)***
Parms	18	15	15	18	15	15
R-squared	0.135	0.240	0.113	0.308	0.345	0.228
Chi-squared	19.55	36.84***	31.25***	55.68***	61.69***	72.47***

***, ** and * denote significance at 1%, 5% and 10% respectively.

Table 11.

MNL for the linkage between working conditions and wellbeing of cocoa farmhands.

Explanatory variable	Coefficient (z-value)
	Living standard: base outcome = low	Resilience: base outcome = low		Food security: base outcome = high	Health: base outcome = high	Asset: base outcome = low
	Moderate	Moderate	High	Moderate	Moderate	Moderate	High
Sex	−0.442 (−0.53)	−10.575 (−0.01)	−10.649 (−0.01)	0.650 (0.78)	15.839 (0.01)	2.247 (1.99)**	13.225 (0.01)
Age	0.001 (0.03)	−0.162 (−1.68)*	−0.154 (−1.56)	0.024 (0.85)	−0.011 (−0.29)	−0.015 (−0.52)	0.036 (0.53)
Education	0.055 (1.61)	−0.219 (−1.41)	−0.210 (−1.34)	−0.016 (−0.51)	−0.076 (−1.80)*	0.063 (1.89)*	0.096 (1.18)
Dependents	−0.092 (−1.36)	0.509 (1.75)*	0.548 (1.87)*	−0.016 (−0.25)	0.154 (1.85)*	0.063 (0.99)	−0.060 (−0.35)
Off farm	0.193 (0.60)	1.272 (1.06)	1.371 (1.12)	−0.017 (−0.06)	0.353 (0.95)	0.892 (2.83)***	0.029 (0.04)
Years as a farmhand	0.004 (0.12)	0.082 (0.60)	0.145 (1.04)	0.017 (0.59)	−0.027 (−0.69)	0.012 (0.39)	0.023 (0.34)
Closeness to social amenities	−0.022 (−4.31)***	0.006 (0.33)	0.008 (0.43)	−0.005 (−1.25)	0.003 (0.57)	−0.016 (−3.63)***	−0.036 (−1.88)*
Working days	0.772 (2.68)***	0.555 (0.71)	0.370 (0.46)	−0.456 (−1.97)**	0.483 (1.49)	0.186 (0.74)	0.992 (2.11)**
Working hours	0.244 (2.11)**	0.443 (0.85)	0.321 (0.61)	−0.239 (−2.22)**	−0.058 (−0.41)	0.007 (0.06)	0.340 (1.32)
Type of contract	0.337 (1.00)	3.177 (2.40)**	4.803 (3.55)***	−0.587 (−1.84)*	−0.067 (−0.17)	0.632 (1.92)*	−0.634 (−0.85)
Category of farmhand	0.969 (2.83)***	−0.475 (−0.40)	−0.727 (−0.60)	−0.239 (−0.76)	−0.607 (−1.59)	0.440 (1.37)	0.039 (0.05)
Number of farmers farmhand serves	−0.036 (−1.35)	−0.001 (−0.01)	0.002 (0.02)	−0.006 (−0.29)	−0.029 (−0.93)	−0.003 (−0.12)	−0.050 (−0.83)
Bonuses	−0.049 (−0.16)	−0.009 (−0.01)	0.060 (0.06)	0.272 (0.96)	−1.311 (−3.36)***	−0.286 (−0.97)	−0.500 (−0.69)
Satisfaction with conditions of work	−11.069 (−0.66)	−70.973 (−1.02)	−70.238 (−1.00)	−2.218 (−0.14)	1.919 (0.09)	−14.171 (−0.88)	10.662 (0.24)
Income from being a cocoa farmhand	0.0002 (0.31)	0.0003 (0.14)	0.0007 (−0.001)	−0.0001 (−2.76)***	0.0002 (0.39)	0.0001 (−0.02)	0.0002 (1.87)*
Constant	3.112 (1.51)	11.378 (0.01)	10.787 (0.01)	−0.943 (−0.52)	−17.018 (−0.01)	−3.878 (−1.86)*	−13.259 (−0.01)
Observations	400	400		400	400	400
Likelihood ratio chi-squared	66.72***	76.42***		28.78**	52.68***	67.20***
Pseudo R-squared	0.198	0.193		0.087	0.187	0.161
Log likelihood	−134.97	−159.98		−151.98	−114.40	−175.75

***, ** and * denote significance at 1%, 5% and 10% respectively.

Sex is significant and negative for migrants for living standards but positive for asset ownership in the SURE (Table 10a, Table 10b). This implies that male migrants are likely to have low living standards than females. However, being a male increases asset ownership. Also, sex is significant and positive for moderate in the MNL for asset ownership (Table 11). This implies that male farmhands are likely to have moderate asset score while females are likely to have low asset score. Male farmhands are more likely to reside in farmsteads than females. Farm settlement reduces access to potable water, electricity and other amenities necessary for living standards. Generally, males migrate as farmhands than females. After fully settling, men/husbands make women/wives join them. Majority of women migrate for marriage motives [2,11]. In such cases, women join men when men have earned enough for better living standard. As a result of the difficult nature of cocoa farm work, men are more likely to have the strength to cater for larger cocoa farms as farmhands than women. This increases the income of male farmhands for purchasing assets than females.

Age is significant and negative for migrants for food security in the SURE. Thus, food security reduces with age. Younger farmhands are better in dietary diversity and access and food availability than older labourers. Younger farmhands are more energetic and can cater for larger cocoa farms or work for many cocoa farmers than older ones. This increases their income to diversify food consumption. Younger farmhands have the extra strength to manage their own farms, which enhances food security. Generally, the older generation are more accustomed with particular diet(s). This reduces their dietary diversity. Also, age is significant and negative for moderate in the MNL for resilience. This suggests that older farmhands are likely to have low resilience to shocks while younger ones are likely to have moderate resilience. Farm ownership, credit accessibility and social capital improve resilience to shocks. With extra strength to own farms, younger farmhands are likely to be less vulnerable to food and income shocks. Being more adventurous and having longer planning horizons, younger farmhands are highly inclined to try new things like different credit sources. Generally, younger people are fond of group formation and membership like youth/fun clubs than older ones.

Education is significant and positive for the aggregate for asset ownership in the SURE. Thus, asset ownership increases with education. Similarly, education is significant and positive for moderate in the MNL for asset ownership. Thus, education increases the probability of moderate asset ownership but reduces the probability of low asset ownership. Also, education is significant and negative for moderate in the MNL for health. This implies that as a farmhands years of education increases, the probability of having moderate health reduces while that of high health increases. Educated cocoa farmers can better access information [33] Education increases capacity to process information, innovation and cognitive ability [33]. These improve knowledge about income generating activities for the purchase of assets and payment of medical bills. Education also improves knowledge about alternative healthcare services.

Dependents is significant and negative for living standard (migrants) and health (native and the aggregate) but positive for resilience (natives and the aggregate) in the SURE (Table 10a, Table 10b). These imply that dependents reduce living standard and health but increase resilience to shocks. Similarly, dependents is significant and positive for moderate and high in the MNL for resilience, and positive for moderate for health (Table 11). These imply that dependents enhance the possibility of having moderate or high resilience but reduces that of low resilience. Also, dependents enhance the possibility of having moderate health but reduces that of high health. Thus, with more dependents, living standard and health reduces but resilience increases. People with more dependents tend to have greater financial burden [48]. This reduces their ability to afford necessities like good housing, toilet and other facilities required for good living standard. Also, financial burden reduces farmhands’ ability to bear medical bills. However, being aware of the financial burden, farmhands having many dependents might seek permanent/long-term contract to ensure job security and reliable income source. They might also do their own farms to obtain extra income and food to feed dependents. These are likely to improve resilience to shocks.

From the SURE, off-farm income boosts health and asset ownership. Likewise, the MNL reveals that off-farm income boosts asset ownership. Farmhands obtain extra income from off-farm jobs. This enhances their ability to afford medical bills and purchase assets. Furthermore, the SURE shows that the number of years a farmhand had been working on cocoa farms has positive association with resilience to shocks and asset ownership. Farmhands who had worked for many years are more likely to accumulate money over time. This enables them to cope better with food and financial crises and purchase assets. Farmhands build trust and networks with cocoa farmers as they work for many years. This influences cocoa farmers’ decision to give them permanent/long-term contract or give them loans in times of financial difficulties. This leads to job security and high credit accessibility. Also, farmhands who work on cocoa farms for many years usually built social networks within the industry. They could seek assistance from people they had built links with.

From the SURE, distance for assessing social amenities negatively influences living standards, health and asset ownership. Likewise, the MNL shows that distance to social amenities negatively influences living standards and asset ownership. Longer distances to social amenities/infrastructure like electricity, schools, clinics, entertainment centres, and motorable roads reduce their usage by farmhands, which reduce living standards and health. Farmhands in communities without electricity might not purchase electrical devices/assets like mobile phones, television, fan, refrigerator, sewing machine and electric stove.

Type of migrant is significant and positive for migrants in the SURE for health and asset ownership (Table 10a, Table 10b). Thus, being a permanent migrant improves health and asset ownership. Compared with temporary/seasonal migrants, permanent migrants had stayed in host communities for longer years. This enables permanent migrants to accumulate ample money to purchase assets and pay medical bills. Also, whether household member had joined migrant is significant and positive for food security (Table 11). This suggests that household members staying with migrants could be adult who assist migrants in undertaking cocoa production activities, own farm and other economic activities. These assistance increases migrants’ income and food availability which enhance food security.

Working days and hours improve living standards, resilience, food security, health, and asset ownership in the SURE. Similarly, the MNL shows that working days and hours improve living standards, food security and asset ownership. Income of farmhands, especially, by-day-workers (casual workers) increases with number of working days and hours. This is because casual farmhands are only paid on days they secure a spontaneous contract to work for a cocoa farmer. For caretakers of cocoa farms (permanent farmhands), working for many days and hours increase cocoa productivity, and subsequently, income. Income enables farmhands to afford the cost associated with improving living standards, resilience to shocks, food security, health, and asset ownership.

From the SURE, nature of contract boosts resilience, food security, health and asset ownership. The SURE also reveals that category of farmhand boost living standard, food security, health, and asset ownership but reduces resilience. From the MNL, nature of contract improves resilience and asset ownership but reduces food security. The MNL further shows that category of farmhand improves living standard. Farmhands with permanent contracts and caretaker-farmhands usually have higher, secured and reliable income sources than those with temporary contracts and by-day/casual farmhands. This increases farmhands with permanent contracts and caretaker-farmhands’ resilience to food and financial shocks, food security, ability to pay medical bills, and asset ownership. The reason why being a permanent farmhand reduces resilience could be that casual farmhands are more likely to have ample time for managing their own farms which increases their resilience to food shocks. Also, the reason why permanent contract reduces food security could be that farmhands with temporary contracts are more likely to have ample time for managing their own farms which increases food availability.

From Table 10a, Table 10b, bonuses enhance resilience and health while income enhances living standards, food security and asset ownership. Similarly, Table 11 reveals that bonuses enhance health while income enhances food security and asset ownership. Income enables farmhands to afford necessities relevant for enhancing their resilience to shocks, ability to afford good healthcare, living standards, food security and asset ownership. Also, satisfaction with conditions of work has positive relationship with health. Satisfaction with conditions of work increases farmhands' commitment to work on cocoa farms. This has a greater possibility of increasing cocoa productivity which subsequently increases farmhands’ income to afford good healthcare. Job satisfaction positively influences productivity [49].

5. Conclusions and implications

Wellbeing (living standards, resilience, health and asset ownership) of migrant and native cocoa farmhands were generally low. For living standards, natives were better off in improve drinking water, electricity, improved sanitation, flooring, building materials and cooking fuel than migrants. Thus, natives had better living standards than migrants. For resilience, migrants were better in terms of credit accessibility, farm ownership and social capital while natives were better in job security. Thus, migrants were more resilient to income and food shocks than natives. For health, natives had better ability to use health facilities than migrants while migrants were better in child mortality than natives. This made the health score of the two the same. For food security, none of the farmhands was deficient/deprived in starchy staples and vegetables. Dietary diversification and access in terms of consuming legumes, meat/fish and fat and oil were generally high while those of fruits, egg/milk and supplementary foods/condiments were low. Majority did not have ample food all-year-round. Migrants were food secured than natives. The mean asset score for farmhands was low, implying low asset ownership. Overall wellbeing of migrants and natives did not differ. They had high in half of the 30 wellbeing indicators/parameters. Working and living conditions like years as a farmhand, distance for assessing social amenities, years in community as a migrant, type of migrant, whether household member had joined migrant, working days, working hours, nature of contract, category of farmhand, bonuses, satisfaction with working conditions, and income from being a farmhand influence living standards, resilience, health and asset ownership. Thus, there is a link between working conditions and wellbeing of cocoa farmhands.

To improve farmhands' living standards, government and private agencies should provide social amenities/infrastructure like potable water, electricity, improved sanitation, health centres and roads in major cocoa-producing communities. Cocoa farmers should assist their farmhands with good accommodation. To improve farmhands’ resilience to shocks, financial institutions should advance soft loans to farmhands. Farmhands should do their own farms, undertake extra income-generating activities, especially, during their off-days (free times), and join associations to augment income and improve social capital. Having their own farms would further improve food security. With the exception of temporary migrants, cocoa farmers should offer farmhands with long-term contracts to ensure job security (reliable income). Farmhands should devote more days and hours to work on cocoa farms to increase productivity. Cocoa farmers should give farmhands bonuses/incentives, PPE and ensure that they are satisfied with working conditions like type of contract, working hours/days, remunerations, welfare/benefits and freedom to motivate them to work hard. It is relevant to sensitize farmhands on the importance of wearing PPE to minimize health/safety risks of cocoa farm activities.

The major limitation of the study is that the issue of workforce safety was not considered. Therefore, it is recommended that future studies on the linkage between working conditions and well-being should consider labour safety as a parameter. Also, migration with children could adversely affect their welfare since those who migrate to work on cocoa farms usually reside in rural areas closer to the cocoa farms. It could also lead to child labour at the destination. However, these were not considered in the study. Thus, future studies should link labour migration to child labour and welfare of children.

Author contribution statement

Bismark Amfo, Ph.D.; Aidoo Robert; Osei Mensah James; Patrick Muotono Izideen Maanikuu: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement

Data will be made available on request.

Declaration of interest's statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.