Gender-based participation in income generating activities in cocoa growing communities. The role of youth training programs.

Despite an impressive growth of Ghana's economy over the decades, it has been branded as 'jobless growth' manifested in growing unemployment among the youth. This growing unemployment rate is gender biased against women, with the United Nations expecting this to worsen globally over time. This study examines the determinants of gender-based participation of the youth in income generating enterprises in cocoa growing areas in Ghana, using a sample of 4,702 participants of the Next Generation Cocoa Youth Program (MASO) initiative. This data were obtained by Solidaridad West Africa. The sample was obtained through the Propensity Score Matching (PSM) technique to compare the characteristics of the male and female groups and remove missing observations. The multinomial regression model supported with descriptive analysis were employed for the estimation. It was found that most of those who participated (63%) in the training program were engaged in an income generating activity. Youth engagement in cocoa (9%) as compared with non-cocoa (28%) and joint enterprise (26%) is very low. Given the same level of exposure (MASO), women are more likely to be unemployed, less likely to engage in cocoa production and more likely to be engaged in non-cocoa enterprises. All non-cocoa business enterprises are dominated by men except petty trading and agro processing. Age, marital status, education, savings, additional training, migrant status and gender, influenced engagements in the various categories of enterprises. The study recommends rolling out other policies that can address challenges of women engagement in cocoa farming in addition to the training program. Policies towards encouraging savings among the youth is recommended to aid in startup businesses, which may be supported by low interest loans. Attention should be given to the non-cocoa sector in terms of employment as people are either losing interest in cocoa farming or diversifying their incomes to non-cocoa businesses.

Introduction

Ghana's unemployment rate has been higher than Sub-Saharan Africa's (SSA) average, which raises a point of concern (Aryeetey and Baah-Boateng, 2015). Despite an impressive economic growth over the decades, it has been branded as ‘jobless growth’ referring to its inability to translate into higher employment especially for the youth (Baah–Boateng, 2013). Unemployment rate in the country currently stands at 8.4% for the total labor force. However, this average figure varies disproportionately among sectors, regions, gender, and age groups (GSS, 2019). Gender and age groups, which are central to this study is further discussed. Unemployment among the female labor force in most recent times has always been higher than their male counterparts (Dadzie et al., 2020). At the current 8.4%, female unemployment rate is 9.2% whiles that of males is 7.5% (GSS, 2019). This phenomenon is also true globally with the United Nation's expectation that gender unemployment disparity will widen (ILO, 2012). This is supported further by empirical study by Baah-Boateng (2013) who found that women are more likely to be unemployed than men in Ghana.

Tracing the pattern of unemployment across gender and age groups from all the Ghana Living Standard Surveys (GLSS) since 1987 confirm the claims above. The detailed tabulation of these is found in Appendix 1. It can be observed that the gender disparity in unemployment was against men in the first two surveys (GLSS 1 & 2). However, the rest of the surveys (GLSS 3,4,5,6 &7) have affirmed that women faces more unemployment than men. This gives rise to the suspicion that factors influencing men and women in their engagement in income generating activities are different. In terms of the unemployment disparity among age groups, it is clear from Appendix 1 that, the youth (15–35 years) have been largely experiencing unemployment among the working labor force. However, its sub group (15–24 years) have consistently taking the largest burden of this unemployment over time. According to GSS (2014), of the largely youthful population of Ghana, this age group makes up 20% of the entire population and most of the unemployed are in this group. Furthermore, 50% of this age group are not in school, 27% are unemployed, and for those who work, 5% work in cocoa farming enterprises (Löwe, 2017).

The above context provides the basis for this study in the sense that, the disparity in terms of unemployment among men and women warrants an investigation into the determinants of their involvements in income generating enterprises. Secondly, the specific age group (15–24) that bears the brunt of unemployment the most is worth targeting. That is why the data used is fit for purpose as far the objectives of this study is concern. The MASO initiative targeted youth from 15 to 25 years in cocoa growing areas with the main purpose of creating employment opportunities (Pinet et al., 2020). Cocoa yields are dwindling with accompanying high unemployment especially among the youth (Löwe, 2017). It is estimated that the average age of cocoa farmers is above 50 years (Oppong, 2015), indicating that, the youth are less involved in cocoa farming. Furthermore, the labor intensive nature of cocoa farming requires young and energetic hands. In light of the above, the main objective of this study is to investigate what determines the level of engagement of men and women in various economic activities after they have all been exposed to the same level of training.

The next generation cocoa youth program (MASO)

One of the ways unemployment among the youth can be tackled is the building of their human capital through market oriented training programs (Kluve et al., 2019). MASO program is just one of the interventions in this regard. At the state level, government has instituted similar tailor made training program for university graduates called The Nation Builders Corps (NABCO). MASO spanned a five-year period (2016–2020) and the data used for the study are responses from the beneficiaries who were trained with the necessary skills to create or join income generating activities, including cocoa farming (Pinet et al., 2020). The MASO initiative has six consortia; Solidaridad West Africa, Aflatoun, Ashesi University, Fidelity Bank Ghana, Opportunity International, and the Ghana Cocoa Board. This initiative aims to develop the interest of youth to quality employment in cocoa and non-cocoa enterprises by creating business opportunities to start their own businesses. To achieve this objective, the MASO project recruited and provide training for the youth to make a career in cocoa production and/or other related enterprises to deal with unemployment in rural Ghana. The MASO model employs a holistic approach to reduce the barriers for the youth to participate in cocoa and non-cocoa enterprises by providing them with important market-oriented skills, training, coaching, mentorship, and creating a healthy environment through youth networking, access to land, finance, and market services (Löwe, 2017).

The MASO program's implementation model is in three main components; Agro Academy, Business Academy, and the Alumni Network (MASO Connect). The first two is the real training whiles the third is just a support for the trainees after they have successfully pass out from any of the two. The Agro Academy focuses on cocoa farming which is the most prevalent farming activities in the area. This training is done with the expectation that trainees will engage in cocoa farming enterprises using sustainable farming practices. The Business Academy aims at developing the skills of the youth outside cocoa farming to other business enterprises that will still support the cocoa industry since it is the main stay of the people in the area. The Alumni Network (MASO Connect) comes in after the training to support the cocoa farmers and other business enterprises in terms of access to finance, land, markets, etc (Löwe, 2017). Alongside, the program has been very sensitive to gender in terms of recruitment and the needs for participants. The “MASO Gender Inclusivity” agenda ensured that the designated Agro and Business Academies addressed specific needs of female participants. Therefore, the gender segregation analysis of this study hinges strongly on the fact that the MASO program deliberately incorporate gender inclusivity in the rollout of its activities and programs (Pinet et al., 2020).

In relation to the MASO initiative, several studies have been done addressing different aspects from this study. Mabe et al. (2020) examined the determinants of youth involvement in cocoa value chain activities by comparing MASO and non-MASO participants. Among other things, the most essential was the participation in MASO, which implies that youth in these communities need training opportunities like this to increase their participation in cocoa value chain activities. Azumah et al. (2021) investigated the relationship between participating in MASO and migration, and found a negative and significant relationship. Youth who participated in the program are less likely to migrate. An impact evaluation study of the initiative was done by Pinet et al. (2020) who found that there is a great potential for the youth who participated in the program. There was an increase positive attitude towards cocoa farming and agriculture in general. Financial inclusion was another impact established by the study. Löwe (2017) considered the program as a great opportunity for youth in cocoa growing area as it served as a link for them to sustainable farming, access to finance, marketing etc. The focus of this study is unique in the sense that, it targets only those that have successfully gone through the program, and try to investigate the determinants influencing their current engagements in income generating activities, which was the sole purpose of the program.

Why cocoa growing communities

Cocoa is an important cash crop that contributes substantially to the Gross Domestic Product (GDP) of several economies in Africa (World Cocoa Foundation and KIT, 2017) by improving their foreign exchange reserves (Danso-Abbeam et al., 2020). The production of cocoa contributes significantly to the country's GDP by 1.8%, while employing over four million people in Ghana (GSS, 2018). In terms of foreign exchange for agricultural produce, the share of cocoa comprises nearly 80% (ISSER, 2017a). Ghana has a huge potential in cocoa production as the country has good environmental advantages as well as suitable climatic conditions (Abdulai et al., 2018). At the micro level, it supports livelihoods of many in Sub-Saharan Africa including Ghana. The cocoa enterprise sector in the economy has a huge potential to assist government and non-governmental organizations in formulating policies to eradicate extreme unemployment, poverty and hunger (Cerda et al., 2014). The sector is known to create sustainable jobs for people in rural areas of developing economies like Ghana (Adesugba and Mavrotas, 2016).

It has been estimated that worldwide, about 5–6 million farmers (both males and females) participate in cocoa production using about 10 million hectares of land (Skalidou, 2020). About 2 million cocoa producers have been estimated to be in West Africa and about 25% of them are females who own cocoa farms (Barrientos, 2013). Due to the huge stake cocoa production has, researchers, international organizations and cocoa producing economies have come to agree that the future for sustainable cocoa production is bleak if the current average age of farmers is maintained, which is estimated to be above 50 years (Laven and Boomsma, 2012; Oppong, 2015; Wessel and Quist-Wessel, 2015; Löwe, 2017; Akrofi-Atitianti et al., 2018; Tsiboe, 2021; Oyekale, 2021). To make matters worse, it has been found that the youth in the cocoa growing area are not interested in cocoa production as a business (Bymolt et al., 2018). This apathy, IDEG (2017) noted “…. has resulted in mass unemployment and lack of sustainable livelihood among youth.” in cocoa growing communities.

Initiatives intended to reverse this trend, has also been criticized for lack of focus. There has been Government's Planting for Food and Jobs (PFJ) program, Cocoa Mass Spraying Program, Cocoa Hand Pollinators program, etc., which are being led by Ghana Cocoa Board (Adego et al., 2019). These interventions are criticized on the basis that they have no focused target group, since their implementations are en masse. The consequences of this approach has been the unintended exclusion of particular groups like women and the youth who do not have equal opportunities especially in the ownership cocoa farm lands (Laven and Boomsma, 2012).

The concept of gender, which also affects participation in cocoa production, is fast evolving and has become of great interest to many researchers and policymakers both at international and local domains (Lambrecht et al., 2018; Yokying and Lambrecht, 2020; Partey et al., 2020). In the context of Ghana, it has been documented that about 20% of females are directly or indirectly engaged in cocoa production (Barrientos and Bobie, 2016). It is therefore observed that females’ participation in cocoa enterprises is significantly lower than males in rural areas. The MASO program has been hailed as a targeted initiative encouraging the youth to engage in both cocoa and non-cocoa enterprises in order to reduce unemployment and promote the welfare of the cocoa growing communities as a whole (Löwe, 2017; Pinet et al., 2020). The FAO (2020) has included the MASO initiative as one of the youth and gender sensitive program capable of increasing agricultural investment and improving the food systems in Ghana. Another initiative targeting the youth in this sector is the establishment of the “Young cocoa farmer award” since 2012 (Oppong, 2015).

Given the same training and exposure, a gender segregated study will help in revealing the importance of gender-based specific determinants responsible for the participation of females and males in cocoa and non-cocoa enterprises and at the same time a good contribution to the literature as well (Wedajo et al., 2019; Nwozor and Olanrewaju, 2020).

Methodology

Study area

Cocoa production is concentrated in the Southern part of Ghana. Hence the study was carried out in six demarcated cocoa regions (Volta, Ashanti, Western North, Central, Oti and Ahafo regions) using 11 districts (see Figure 1) - where the MASO program was implemented. The data is a product of a five year (2016–2020) initiative. The areas are characterized as forested and have a good ecosystem for cocoa production. The production of cocoa occurs in two seasons, minor and major seasons, as rainfall in these areas is bi-modal and evenly distributed than other parts of the country.

Figure 1

Map of Ghana showing the MASO program districts. Source: Authors' Construct from field data, 2020.

Sample

The study used data from the MASO program collected by Solidaridad and partner organizations. The MASO program recruited nearly 12,000 youth, both females and males and trained about 70% of them in business and entrepreneurship in the cocoa and non-cocoa value chain activities. For each of the interviews conducted with the participants, a consent form was presented to each respondent by enumerators for acceptance before administering the questionnaires. Since gender is the focus of this study, PSM was used to compare the characteristics of male and female groups within the data. The matching showed that using gender as the treatment, the two groups are similar and would be free from bias and confoundedness, when comparison is made between their covariates (Austin, 2011). Furthermore, the technique also helped to remove missing data. The lost data points after implementing the PSM command means that there were observations on the dependent variable under investigation which needed to be dropped for further analysis. Based on the variables of interest (i.e. gender, cocoa and non-cocoa enterprise), a total of 4,702 observations (data points) were generated and used for the analysis. This data points are made of 2655 males (56.5%) and 2047 females (43.5%).

Analytical framework

This study is underpinned by the random utility maximization theory since the participation in cocoa and non-cocoa enterprise portfolios by females and males is based on the satisfaction, they obtain rather than not engaged in any economic activity. In this study, business, enterprise or business enterprises are used interchangeably, and refereeing to any form of income generating activity an individual engages in. Females and males in cocoa communities engage in different kinds of enterprises ranging from cocoa, food crop production, other cash crop production, animal production, petty trading, mobile-money (Momo) vendors, artisanship, agro-food processing, mining enterprises etc.

Using the implementation model of the MASO program, the study categorizes the respondents into four (4); those who are engaged in only cocoa farming enterprises as the result of the Agro Academy, those engaged in other enterprises as the results of the Business Academy, those engaged in both cocoa farming and other enterprises after the training, and those who are not engaged in any enterprises after the training. The third category emphasizes the livelihood diversification of people in these areas, where cocoa farmers engages in other non-cocoa activities simultaneously. Each of these categorizations is further disaggregated into male and female.

Several studies used different models to examine the drivers of smallholders' decision or choice behavior of technologies and the participation in agricultural income generation activities. Where more than two categorical dependent variables exist, quantitative models like multivariate logit, probit, ordered logit/probit, and multinomial logit/probit models are mostly used (Addisu et al., 2016, Gebrehiwot and Van Der Veen, 2013, Teklewold et al., 2013). According to Ben-Akiva et al. (1985), the simple logit and probit model estimation is built on the behavioral choice decision in relation to participation in an economic activity like cocoa and non-cocoa enterprises. When the outcome variable is solely single and binary, logit or probit models are appropriate. Also, when the outcome variable is categorical (choice among many options in an orderly manner), then ordered probit is the best approach. However, when a rational economic agent decides to participate or not to participate in any economic activity as aforementioned based on the economic agent's main decision, then probit/logit or ordered probit fails to handle such a situation, justifying the choice of multinomial logit/probit modelling. The merits of multinomial logit/probit models have been well discussed by Tabachnick et al. (2007) as the best alternative to modelling approaches for categorical variable outcomes. This study employed the multinomial probit model to examine the determinants of participation of females and males in cocoa and non-cocoa enterprises owing to the fact that they have received training through the MASO program. The data were segregated by gender. The multinomial probit regression model was then used to analyze the data for both females and males separately.

The modelling of the choices each participant makes is based on the assumption that each participant is exposed to the various categories of enterprise engagements. Structurally, the multinomial probit model is given as; is a categorically-distributed data, where each outcome j for each observation i happens under a probability which is unobserved. That is, the expected utility a participant benefit from choosing a particular jth enterprise category for each observation is latent. The represent the vector of variables that are suspected to have an influence on the participants choice of enterprise. The and are the coefficients and the error term respectively. The error term are assumed to follow a multivariate normal (Gaussian) distribution with a mean zero (0) and covariance matrix of . The base category (the category to which all other categories are compared) in the matrix would be normalized to one in the main diagonal (Greene, 2012).

As indicated in Eqn 1, the expected utility () participants get from their decisions to engage in any business enterprise is not directly observed, but measured as a latent variable to define the choice categories. Eqn 2 acts as the link between the latent and the observed, such that category j (j = 1,2,3……m) is chosen only when is highest for all other j categories. is the choice category under the observed situation.

Under utility maximization, the choice of a category j would mean;

The probability of occurrence of this event in Eqn 3 is translated into;

The coefficients resulting from Eqn 4 indicates the direction and the likelihood of the outcome category happening. However, the marginal effect gives more information about the relationship by providing the exact probabilities in addition to the direction of the effect (Wulff, 2015). The marginal effect equation for a multinomial probit model is given in Eqn 5 as:where is the probability weighted average of the coefficients for the various choice outcome combinations (Wulff, 2015).

In this study, the 4 j-categories are (J1); those who could not be engaged at the time of the interview, but has successfully participated in the training program, (J2); engagement in non-cocoa enterprise only, (J3); engagement in cocoa enterprise only, and (J4); engagement in joint business enterprise (cocoa and non-cocoa enterprises). J1 is the base category. The definitions of Xs and their expected signs are presented in Table 1.

Table 1

Definitions of variables and their expected signs (Gender segregated and pooled data).

Factor	Definition of factor	Expected sign (Gender segregated and pooled)
		Non-cocoa enterprise	Cocoa enterprise	joint enterprises
No business at all	Youth did not engage in any business at all; 0	N/A	N/A	N/A
Non-cocoa business	Youth engaged in only non-cocoa business; 1	N/A	N/A	N/A
Cocoa business	Youth engaged in only cocoa business; 2	N/A	N/A	N/A
Joint business	Youth engaged in both cocoa and non-cocoa business; 3	N/A	N/A	N/A
Savings	If youth engaged in savings at a financial institution (1/0)	+	+	_
Education	The educational level of a youth in years	+/-	+/-	+
Migration	If youth is a migrant (1/0)	-	-	-
Marriage	If youth married (1/0)	+/-	+/-	+/-
Mobile-phone	If youth own a mobile-phone (1/0)	+	+	+
Advocacy group	If youth is part of a community development service group (1/0)	+	+	+
Leadership	If youth in any leadership position in any community service group (1/0)	+	+	_
Training	If youth gained additional training in cocoa and non-cocoa enterprise (1/0)	+	+	+
Information transfer	If youth share the knowledge learned in MASO program (1/0)	+	+	+

Source: Authors' construct, 2020.

Results and discussions

Descriptive statistics of respondents

The summary of gender segregated statistics as well as the pooled are presented in Table 2. The results reveal that, averagely for all those involved in the training, 37% could not find job at the time of the data collection. However, this proportion is less for men (33%) than women (43%). Even though both genders were given the same opportunity in terms of the training, women still lag behind men in finding opportunities for economic activities. In the non-cocoa enterprise engagement, 28% of the participants were engaged, however women almost doubled the proportion of men in this sector; males (20%) and female (38%). Conversely the proportion of men engaged in cocoa enterprises is about the triple that of females. Even though the proportion (9%) is low, it appears the opportunity for males in the cocoa sector is more than females even after exposing them to the same training program. As high as 26% of the participants were engaged in both cocoa enterprises as well as at least one non-cocoa business enterprise. This gives an indication of diversification of the income sources of the youth in cocoa growing communities (Knudsen and Agergaard, 2015). At the gender level, men (34%) are found to be more engaged in this diversification drive than women (15%).

Table 2

Descriptive statistics.

Indicator	Pooled		Gender				Chi2-test
			Male		Female
	Freq.	Percent	Freq.	Percent	Freq.	Percent
No enterprise engagement	1745	37	868	33	877	43	420.797∗∗∗
Non-cocoa enterprises	1304	28	532	20	772	38
Cocoa farming enterprises	443	9	354	13	89	4
Both cocoa and non-cocoa enterprises	1210	26	901	34	309	15
Total	4702	100	2655	100	2047	100
Marriage	1619	34.43	782	29.45	832	40.64	66.948∗∗∗
Additional-training	880	18.72	525	19.77	355	17.34	4.492∗∗
Information transfer	550	11.70	403	15.18	147	7.18	71.577∗∗∗
Savings	3765	80.07	2186	82.34	1579	77.14	19.572∗∗∗
Leadership	4245	90.28	2376	89.49	1869	91.30	4.329∗∗
Advocacy group	643	13.68	432	16.27	213	10.41	34.82∗∗∗
Mobile-phone	4218	89.71	2518	94.84	1700	83.05	174.043∗∗∗
Migration	1458	31.01	899	33.86	559	27.31	23.197∗∗∗
Age (mean)	22.2		22.32		22.05		25.452∗∗∗
Education (mean)	9.88		10.44		9.16		372.177∗∗∗

∗∗∗p < 0.01, ∗∗p < 0.05.

The Chi2 test confirmed that there is a statistical difference between females and males when it comes to enterprises engagements in cocoa and non-cocoa among the youth in the cocoa growing regions. This finding is consistent with the studies by Le-Baron and Gore (2020), which indicated that gender dynamics play a significant role in cocoa farming enterprises. Also, this finding is similar to that of Djokoto et al. (2016), which highlighted that gender is a key factor in the participation or otherwise in cocoa farming in Ghana.

Turning to the socioeconomic indicators, about 29.5% and 40.6% males and females were married respectively. After participation in the MASO program, some participants gained additional training to boost their knowledge and skills in entrepreneurship, cocoa and non-cocoa enterprises. Even though the proportion that got the opportunity was low (18.7%), the difference between both gender is about 2%, meaning that those additional training programs are available to both genders. Information transfer among beneficiaries of training program is expected to have an influence on their ability to be engaged in economic activities. About 15.2% of males and 7.2% of females shared skills and knowledge they acquired with other people in their various communities. The ability to save money is expected to boost the chances of participants to engage in economic activities through easy access to starting capital or loans. The saving habit among participants were high as about 80.1% and 77.1% of males and females respectively were saving their surplus income in various financial institutions. The reason for such finding could be that females mostly spent most of their income on household expenses which reduced their ability to save. Most of the participants had a taste of leadership roles either in the training itself or in their engagements at the community level. The community level is mostly tie to their involvement in community advocacy groups which is dominated by women from Table 2. It was a deliberate part of the training program that participants should be exposed to leadership roles. About 90% of all participants had some leadership roles to play, with almost equal proportion among the genders. Mobile phone usage is expected to have an influence on individual's social and economic life, especially in the age of Information Communication Technology (ICT). Most of the participants (90%) had mobiles phones with the male participants ahead of their female counterparts. Cocoa growing areas have historically been an attraction for labor especially from northern Ghana (Knudsen and Agergaard, 2015). About a third of the participants are migrants, out of which men dominated women.

In terms of age, the average of the participants is 22 years with both genders averaging the same. The average school going years for all the participants is 10, which indicates all of them had at least completed Junior High School. The Chi2 test revealed that all the socioeconomic indicators for males and females were statistically different from zero. This implies that all the socioeconomic factors which will be used in the analysis (segregated and pooled) play distinct roles among males and female participants in the study.

Gender dimension of non-cocoa business enterprises

Aside cocoa farming and its related enterprises, youth in the communities also involve in non-cocoa economic activities. Table 3 presents the results of cross-tabulation of gender and non-cocoa enterprises in cocoa growing communities in Ghana. This analysis was performed using a descriptive statistics approach and the Chi2 test to examine the relationship between gender and each of the enterprises. The total sample size of the participants engaging in non-cocoa enterprises is 2514, which constitute about 53% of the participants. This includes those engaging in only non-cocoa enterprises (1304) and those combining it with cocoa enterprises (1210). Of this sub-sample, men represent 1081(43%) whiles women are 1433 (57%).

Table 3

Cross-tabulation of gender and non-cocoa enterprises.

Non-Cocoa Business	Female		Male		Pooled		Chi2 test
	Freq.	Percent	Freq.	Percent	Freq.	Percent
Food crops
No	882	82	711	50	1593	63	Chi2(1) = 271.382∗∗∗
Yes	199	18	722	50	921	37
Total	1081	100	1433	100	2514	100
Cash crops
No	1067	99	1222	85	2289	91	Chi2(1) = 136.3681∗∗∗
Yes	14	1	211	15	225	9
Total	1081	100	1433	100	2514	100
Petty trade
No	275	25	1285	90	1560	62	Chi2(1) = 1.1e+03∗∗∗
Yes	806	75	148	10	954	38
Total	1081	100	1433	100	2514	100
Transportation
No	1079	100	1212	85	2291	91	Chi2(1) = 176.977∗∗∗
Yes	2	0	221	15	223	9
Total	1081	100	1433	100	2514	100
Government worker
No	1077	100	1252	87	2329	93	Chi2(1) = 135.874∗∗∗
Yes	4	0	181	13	185	7
Total	1081	100	1433	100	2514	100
Construction
No	1076	100	1224	85	2300	91	Chi2(1) = 157.799∗∗∗
Yes	5	0	209	15	214	9
Total	1081	100	1433	100	2514	100
Mobilemoney
No	1057	98	1284	90	2341	93	Chi2(1) = 64.305∗∗∗
Yes	24	2	149	10	173	7
Total	1081	100	1433	100	2514	100
Artisan
No	974	90	1142	80	2116	84	Chi2(1) = 50.1∗∗∗
Yes	107	10	291	20	398	16
Total	1081	100	1433	100	2514	100
Apprenticeship
No	1012	94	1186	83	2198	87	Chi2(1) = 66.049∗∗∗
Yes	69	6	247	17	316	13
Total	1081	100	1433	100	2514	100
Agroprocessing					0	0
No	929	86	1378	96	2307	92	Chi2(1) = 85.226∗∗∗
Yes	152	14	55	4	207	8
Total	1081	100	1433	100	2514	100
Animal rearing
No	1073	99	1359	95	2432	97	Chi2(1) = 38.219∗∗∗
Yes	8	1	74	5	82	3
Total	1081	100	1433	100		100

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

It can be observed from Table 3 that petty trading engages most of the youth (38%). This activity normally involves buying and selling mostly to the final consumers of the product. They are normally at the end of the distribution chain of these consumables and household items. At the gender level, 75% of women are engaged in this economic activity as compared with only 10% of men. A study by Peprah et al. (2019) in Ashanti region (one of the regions in this study) found that women controls the petty trading activities in the region. However, Tufuor et al. (2016) added that income from petty trading is low as compare to other income generating activities. The studies of Ojo et al. (2015) and that of Tu Huynh (2016) further confirms this finding by indicating that the proportion of females participating in petty trading are higher than their male counterparts. Furthermore, it is reported by FAO and ECOWAS (2018) that even though women dominate this aspect of business activities, the dominant focus of the trading is agricultural output of mostly non-cash crops like roots and tubers, cereals, legumes, fruit and vegetables. The trading and marketing of cash crops like cocoa and rubber are controlled mostly by men. Chi2 value was found to be highly significant at 1% level.

Food crop farming takes the second position in terms of engagements in non-cocoa enterprises. About 37% of those engaged in non-cocoa activities are into food crop production. The food crops are mainly for consumption as the cocoa farming are purely for cash. Men dominates this activity with 50% as compared with 18% for women. Since it is for the sustenance of the family, it is normally seen as the duty of the men to own a farming. Conversely, cash crop production only engages 9%, with a very wide gender disparity in favor of men. The national agricultural gender profiling by FAO and ECOWAS (2018) reveals a drastic shift of men to the production of cash crops forcing the women to diversify into food crop production for the sustenance of the household. The proportion is low because cocoa as the main cash crop is not included. There are other cash crops such as rubber, coffee, cashew etc. in these areas but compared with cocoa, they are less cultivated. The next are artisanship and apprenticeship which comprise 16% and 13% of the non-cocoa engagement respectively. Whiles artisanship is a permanent income generating activity, apprenticeship are means to a permanent job. However, it is also an income generating activity for those engaged as it sometimes takes years to complete. Women engagement in both artisanship and apprenticeship is low compared to men.

The rest of the activities recorded lower engagements (below 10%) of the participants. However, with the exception of agro processing, men dominate all the other enterprises. The disparity is very obvious in transportation, government work, and construction where no female among the participants were engaged. Statistically, there was a strong relationship between gender and all the non-cocoa enterprises the participants were engaged in. This is indicated by the 1% significance level.

Determinants of participation in cocoa and non-cocoa enterprises

The estimates of the multinomial regression model are presented in Table 4. Even though the base category is the group without any engagement (No enterprise engagement), the reported results are the marginal effects which gives both the direction and the probabilities. The Multinomial probit model is well behaved as the Wald Chi2 test for females (Chi2 = 377.604; P > χ2 0.000), males (Chi2 = 490.646; P > χ2 0.000) and the pooled (Chi2 = 1166.88; P > χ2 0.000) were highly statistically significant at 1%. These lead to the rejection of the null hypothesis that the coefficients of all variables are jointly equal to zero.

Table 4

Marginal Effects of the Multinomial regression estimates for gendered participation in cocoa and non-cocoa.

Variable	Women		Men		Pooled
	Coeff	Std. Err.	Coeff	Std. Err.	Coeff	Std. Err.
No enterprise engagement
Age	-0.034∗∗∗	0.004	-0.038∗∗∗	0.003	-0.037∗∗∗	0.003
Marriage	-0.152∗∗	0.020	-0.190∗∗∗	0.020	-0.173∗∗∗	0.014
Education	0.008∗	0.004	0.014∗∗∗	0.003	0.011∗∗∗	0.003
Information transfer	-0.138∗∗∗	0.042	-0.109∗∗∗	0.026	-0.116∗∗∗	0.022
Savings	-0.221∗∗∗	0.023	-0.109∗∗∗	0.021	-0.159∗∗∗	0.016
Training	0.153∗∗∗	0.026	0.103∗∗∗	0.021	0.124∗∗∗	0.016
Leadership	-0.110	0.095	-0.038	0.047	-0.050	0.042
Advocacy group	-0.134	0.089	-0.078∗	0.040	-0.084∗∗	0.037
Mobile-phone	0.021	0.028	0.019	0.039	0.016	0.022
Migration	-0.027	0.023	-0.049∗∗∗	0.018	-0.039∗∗∗	0.014
Gender					-0.108∗∗∗	0.013
Non-cocoa enterprise
Age	0.010∗∗	0.004	-0.003	0.003	0.003	0.003
Marriage	0.073∗∗∗	0.021	0.040∗∗	0.017	0.056∗∗∗	0.013
Education	0.006	0.004	0.004	0.003	0.005∗	0.002
Information transfer	0.125∗∗∗	0.040	-0.028	0.023	0.021	0.021
Savings	0.172∗∗∗	0.026	-0.026	0.020	0.061∗∗∗	0.016
Training	-0.111∗∗∗	0.029	-0.027	0.020	-0.064∗∗∗	0.017
Leadership	-0.009	0.092	-0.036	0.042	-0.038	0.040
Advocacy group	0.017	0.086	-0.039	0.036	-0.033	0.036
Mobile-phone	-0.006	0.029	0.012	0.037	0.010	0.022
Migration	0.027	0.023	0.017	0.016	0.020	0.014
Gender					-0.173∗∗∗	0.012
Cocoa farming enterprises
Age	0.001	0.002	0.010∗∗∗	0.003	0.006∗∗∗	0.002
Marriage	0.006	0.009	-0.022	0.014	-0.008	0.009
Education	-0.006∗∗∗	0.002	-0.016∗∗∗	0.002	-0.011∗∗∗	0.001
Information transfer	0.012	0.017	0.025	0.018	0.022∗	0.013
Savings	0.011	0.011	0.033∗	0.018	0.023∗∗	0.011
Training	-0.010	0.014	-0.028	0.017	-0.019	0.011
Leadership	0.018	0.042	0.007	0.031	0.016	0.022
Advocacy group	-0.003	0.038	0.073∗∗∗	0.027	0.047∗∗	0.019
Mobile-phone	-0.019∗	0.011	-0.014	0.028	-0.026∗	0.014
Migration	0.005	0.010	0.014	0.014	0.011	0.009
Gender					0.099∗∗∗	0.010
Cocoa and non-cocoa farming enterprise
Age	0.023∗∗∗	0.003	0.031∗∗∗	0.004	0.028∗∗∗	0.003
Marriage	0.073∗∗∗	0.015	0.173∗∗∗	0.018	0.125∗∗∗	0.012
Education	-0.008∗∗∗	0.003	-0.003	0.003	-0.005∗∗	0.002
Information transfer	0.002	0.029	0.113∗∗∗	0.025	0.073∗∗∗	0.018
Savings	0.038∗	0.02	0.102∗∗∗	0.024	0.075∗∗∗	0.016
Training	-0.033	0.023	-0.048∗∗	0.023	-0.041∗∗	0.016
Leadership	0.101∗	0.057	0.066	0.045	0.072∗∗	0.033
Advocacy group	0.119∗∗	0.052	0.045	0.038	0.070∗∗	0.029
Mobile-phone	0.005	0.021	-0.016	0.041	000012	0.022
Migration	-0.005	0.017	0.018	0.018	0.008	0.013
Gender					0.182∗∗∗	0.012
Model summary
Number of obs	2047		2655		4702
Prob > chi2	0.000		0.000		0.000

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

The results showed that age has a positive and significant effect on enterprise engagements. It can be observed that age has a negative relationship with the chances of participants not being engaged in any economic activities. This means that the older a participant gets, the less likely they are to be jobless. This is true for both genders and the pooled as coefficients are significant at 1%. However, for non-cocoa enterprises, the coefficients are positive for only females, indicating that as women ages they are more likely to be engaged in non-cocoa enterprises like the ones discussed earlier. For cocoa farming enterprises, it is generally found that age is positively related to engagements in cocoa enterprises. However, this effect is not significant for females but highly significant for males. This means that as participants grow older they are more likely to be engaged in cocoa enterprises. In the case of those who combines cocoa enterprises with non-cocoa enterprises, age is found to be directly related to the probability of engaging in both enterprises. This means that older people are more likely to diversify their sources of income. Oyekale (2021) and Wessel and Quist-Wessel (2015) have supported the finding that cocoa and its related activities are mostly done by older people within the cocoa growing communities. This explains the finding that older people are less likely to be unemployed in the communities, given an indication that, at least they will have a cocoa farm as they age. It also directly supports the finding that cocoa farming enterprises are likely to be done by older people, but that is not so for females in the study. However, this could be understood from the fact women owns about 25% of cocoa farms in Ghana in addition to the perception that cocoa farming is not for women (Danso-Abbeam et al., 2020). Implying from the finding that older people in cocoa growing areas are more likely to cultivate cocoa than the younger ones, it therefore means that age is more likely to influence engagement in both enterprises as found in the study. This is because Abdulai et al. (2018) found income diversification to be rampant among cocoa farmers because of the seasonal nature of its activities. Furthermore, as people age, their demands and capital expenditure increase, which propel them to diversify their sources of income. Meanwhile, the finding that associates increased age with cocoa enterprises is a worry to many researchers, who think that production and productivity could increase if young and energetic youth are involved in it (Anyidoho et al., 2012). Similarly, Mabe et al. (2020) found age to have a significant positive effect on youth participation in cocoa value chain activities in Ghana.

The marital status of participants is found to reduce the chances of not being engaged in any economic activity in the study area. This is true for both males and female. Unlike unmarried people, married ones are forced to engage in economic activities to cater for the additional responsibilities that come with being married. Furthermore, couples are more likely to be financially sound before marriage (Karney, 2021). According to the finding, the marital status of participants has no significant influence on their engagement with non-cocoa enterprises. It is however positive for only women. This indicates that married females are more likely to participate in non-cocoa enterprises instead of being redundant. This makes sense since spouses and their children could support themselves in monetary terms and also rely on family labor to expand enterprises. For cocoa enterprise engagement, marital status has no influence either in the pooled or the gender segregated data. It however has a positive and significant effect on participants' engagement jointly in cocoa and non-cocoa enterprises. This is true for males, females and the pooled data. That is, men and women who are married are more likely to participate in joint enterprises ventures than singles or those who never married. Having many mouths to feed as a married person pushes for diversification of income; the joint business enterprises. Practicing joint enterprise ventures requires more farm labor which is more accessible by married couple than unmarried (Garner & O'Campos, 2014). When hired labor is too expensive, such people turn to depend on family labor.

Education contrary to aprior expectation is responsible for the economic inactivity of the participants. At the gender level, the story is the same; the more males and female spent time in school the more likely they are to be jobless even after going through the training program. However, the educational profile of the participants indicates that they are mostly junior and senior high school graduates. Currently it takes some time for university graduates to secure or establish their own business, not to talk of junior and senior high school graduates. According to ISSER (2017b), in the first year graduating from universities, only 10% are able to get jobs. The finding is therefore not surprising considering the educational profile of the sample under study. Gender wise, education has no statistical influence on youth engagement in non-cocoa enterprises. On youth engagement in cocoa enterprises education is found to negatively influence their participation. This means that educated youth are more likely not to be engaged in the cocoa enterprises. This finding is true for both male and females. Anyidoho et al. (2012) indicated that there is a diminishing interest in cocoa farming among youth with higher levels of education. Contrary, Mabe et al. (2020) found that people with higher levels of education have the aspiration of owning a farm and hence, are more likely to participate in cocoa production. At 1% significance, education negatively influence the probability of the youth engaging in both cocoa and non-cocoa enterprises. The findings suggested that more years in formal education will reduce the chances of participants to engage in joint enterprises. More years in education expose people to learn strategic knowledge and skills which enhances their skills to diversify their sources of livelihood. However, this finding is true for only women and that of men is statistically insignificant.

Information transfer is another important indicator that influences youth participation in any economic activity. The study found that participants who share and transfer knowledge learnt from the training are less likely to be unemployed. Information sharing increases the networks of individuals which further increase their chances of being economically engaged. The study found that information transfer had a positive significant effect only on females’ participation in non-cocoa enterprises at 1% level of significance. That is to say females who transferred information that was taught during the MASO program have higher probability of participating in non-cocoa business. Females most often are known to share skills and knowledge they learn with their colleagues due to their flexibility of meeting frequently. The sharing of ideas could influence other females to start businesses to generate income. Females hearing their colleagues' success stories in business inspires them to do the same to generate income. In the case of cocoa and joint enterprises, information transfer is found to be positive and significant. This means that those who share information with other people about agribusiness are more likely to participate in cocoa and non-cocoa enterprises jointly. Farmers share agricultural production knowledge and skills among themselves and this assists them to diversify their source of livelihood. The risk associated with agricultural production makes farmers depend on other farmers who have the opportunity to attend agricultural production seminars to acquire knowledge and skills to enhance their production operations.

Irrespective of the gender, the ability to save money by participants decreases their chances of being unemployed after the training. The effect is the same for non-cocoa, cocoa and joint enterprise engagements. This means that participants who engage in savings are more likely to participate in non-cocoa, cocoa and joint enterprises. Savings have the potential to influence people's access to production and/or start-up capital/credit. It acts as a guarantee for access to loans for the establishment or expansion of existing businesses. In the case of cocoa enterprises, this can be used to purchase farm inputs like weedicides/herbicides, fertilizers, insecticides, farm machines and adoption of other capital-intensive technologies. It can also be used to hire labor and farm managers to work on the cocoa farm, which assists in the creation of jobs in rural settings. This finding goes in line with the studies by Alhassan et al. (2020), which indicated that credit access is key in the participation of individuals in agricultural activities. For non-cocoa enterprises, various inputs and necessary investments can be done with these capital. The surplus incomes saved by females and males might be used to invest in new businesses with the aim of making profits, which explains why it having the same effect on the joint cocoa and non-cocoa enterprise engagement.

Extra training after MASO participation is expected to boost the chances of engaging in income generation ventures and at the same time reduce the chances of being unemployed. However, the result has been counterintuitive in the sense that, participants who had additional training are seen to have a higher probability of being jobless. The result is no different at the gender level. In terms of the categorized enterprise engagements, additional training is found to reduce the probability of participants engaging in cocoa, non-cocoa as well as the combination of both. However, that of cocoa is found not to be significant. Despite this, it is understood from the data that those embarking on additional training after participating in MASO are those who are still searching for jobs, with the hope of it boosting their chances. That means for those that got engaged, these additional training were not important to them. This could explain this counterintuitive finding about additional training programs.

The leadership role participants played in the programs and in their various communities does not have any effect on their participation in economic activities, except for those who combined cocoa and non-cocoa enterprises. It is found that involvement in leadership roles increases the chances of being engaged in both enterprises. But it is true for only females. Furthermore, the study found that participants who are engaged in various community advocacy groups are less likely to be unengaged. It plays no role in the engagement of participants in non-cocoa enterprises. However, in the case of cocoa enterprises, it is found to increase the probability of participants getting engaged. For participants that combined cocoa and non-cocoa enterprises, joining an advocacy group played a positive role in their engagement. Even cocoa and non-cocoa engagements tell the same story, but differ at the gender level. In the former, the effect of joining an advocacy group is positive for only males, while in the latter, it is positive for only females. Ownership and use of mobile phones has no significant effect on all the enterprise categories except the cocoa enterprises, which recorded a negative effect. This indicates that participants with mobile phones are less likely to be engaged in cocoa enterprises. This is true for only females as it has no significant effect on males.

The migrant status of participants could only explain why participants are jobless after the training. It is found that being a migrant to the cocoa community reduces the chances of being jobless or increases the chances of engaging in a business enterprise. The finding is the same for both males and females. Economic migrants by their nature are pushed from their original location because of lack or inadequate economic opportunities. In a study in Ghana, Abukari & Al-hassan (2017) noted that most migrants especially from the northern part of the country to the southern cities (cocoa growing areas) are job seekers. Since job seeking is the reason for migrating, they will settle at a place they can get engaged, otherwise they keep moving. In this sense, migrants within the study are not expected to be jobless as the study revealed. However, for all the categories of enterprises, migration status of a participants has an insignificant effect. Since the earlier finding suggest that they are more likely to be engaged in an enterprise than being jobless, it was expected that, they would be engaged in cocoa related enterprises or non-cocoa or both. Knudsen and Agergaard (2015) has acknowledged the fact that migrants' status in Ghana's cocoa growing has been undergoing some form of transformation. They found that between 1966-1979, the sources of income for migrants in Western Region (one of the study regions) was 100% from cocoa farming, but has reduced to 17% as at 2006. Access to land by these migrants has also drastically reduced limiting their engagement in cocoa and non-cocoa enterprises.

The results from the pooled data have been discussed alongside the gender segregated data. However, the effect of gender on the various enterprise engagements is reflected in the pooled data were the variable gender is considered as an explanatory variable. The resultant coefficient is significant for all the categories. It is found that men compared to women are less likely to be unengaged in any economic activity. This is revealed by the negative significant coefficient of the gender variable under no enterprise engagement. However, this finding can be challenged with the ongoing debate on quantifying women participation in economic activities and the economy as a whole. Most women's economic activities are considered hidden on the bases that they are unpaid (Himmelweit, 2002). In the consideration of the economic activities, the focus is normally on the main stream paid activities which normally exclude much of the activities women are involved in. Until there is a real effort to commercialized domestic skills, women would always be left of out in reporting participation in economic activities (Akyeampong and Fofack, 2014). In light of this, it can be observed that the activities considered as economic activities aside the cocoa farming do not include the domestic and unpaid economic activities mostly engaged by women. Out of the eleven (11) enterprise engagements in Table 3, female could dominate in only two; petty trading and agro processing.

In the case of cocoa and non-cocoa engagements, men are more likely to be engaged in cocoa farming enterprises than women, while the women are more likely to engage in non-cocoa enterprises. Being a male increases the probability of engaging in cocoa farms, which could partly be explained by the fact that the total cocoa farm ownership is dominated by men (75%) (Danso-Abbeam et al., 2020). However, Marston (2016) found that there is a serious underreporting of women engagement in cocoa production. This is attributed to the definition of cocoa farms based on land ownership of which women in Ghana are at a disadvantage. Access to land for production has been challenging for Ghanaian women (Aasoglenang et al., 2013; Danso-Abbeam et al., 2020) and has been the main obstacle to their participation in cocoa farming. Hiscox and Goldstein (2014) noted that only 2% of cocoa lands are owned by women. Female cocoa farmers earn 30% less than their male counterparts (Hiscox and Goldstein, 2014), making it difficult for them to access inputs which are costly. Other challenges include adequate access to labor and other inputs, credit access, technical training, high production cost because of hired labor, and having less control and decision making on cocoa revenue (Hiscox and Goldstein, 2014; Bymolt et al., 2018). The finding that demonstrates the higher likelihood of women engaging in non-cocoa enterprises is a reflection from the previous finding. The challenges identified that limit their participation in the cocoa production push them to engage more in non-cocoa business enterprises. Unlike non-cocoa enterprises, access to land for those enterprises are more accessible because cocoa farms require large arable lands acquired over a long period, considering the life time of a cocoa plantation. Results from Table 3 also support this finding as the most prevalent non-cocoa enterprise (petty trade) is dominated by women.

In combining both cocoa and non-cocoa business enterprises, it is found that men are more likely than women to be engaged in this category of business. Cocoa farmers are known to diversify their income by engaging in various non-cocoa enterprises (Djokoto et al., 2017; Abdulai et al., 2018; Amfo and Ali, 2020). Since men are more likely to be engaged in cocoa farming enterprises, it therefore means that they are more likely to engage in other non-cocoa businesses, hence this finding. Following the deliberate promotion of youth participation in cocoa farming, these findings suggest that, despite the encouragement of young women to participate in cocoa farming, they are still lagging behind even if they are given the same form of training as their male counterparts. This means that though there may be unintended gender related benefits from the program (Pinet et al., 2020), men still have more chances of participating in cocoa farming then women.

Conclusion and recommendations

The study sets out to investigate what determines the participation in business enterprises after the MASO initiative. 37% of the participants could not get engaged in any form business activity. This goes to emphasize the positive effect of MASO since majority (63%) of participants were engaged after the training program. It was found that, out of those engaged, cocoa enterprises recorded the least (9%) after non-cocoa (28%) and joint enterprises (26%). This adds to observations by some researchers that economic activities in the cocoa growing frontier is shifting away cocoa production to diversification into non-cocoa activities. However, given the same level of exposure like the MASO program, women are more likely to be unemployed, less likely to engage in cocoa production and more likely to be engaged in non-cocoa enterprises. Aside cocoa production, the next agricultural engagement is food crop farming as opposed to cash crops. This means that other cash crops like coffee, rubber, cashew etc. are not given much attention. Unlike food crops, women engagement in cash crops is negligible. Out of the 11 non-cocoa business enterprise, petty trade and agro processing are the women dominated enterprises. The rest of the non-cocoa business enterprises are dominated by men.

In general, older participants are found to be more likely to be engaged in cocoa farming business. Changing the narrative of older cocoa farmers seem not have changed even if a holistic training and capacity building program like MASO is implemented. In line with this trend, the study found that older participants are less likely to be unemployed. Married youth are less likely to be unemployed but are more likely to be engaged in non-cocoa enterprises irrespective of their gender. Level of education reduces the chances of a youth to be jobless, increases the chance of engaging in non-cocoa enterprises, and rather decreases the probability of participating in cocoa farming as well as diversification of income sources. Aside non-cocoa enterprises where gender segregation does not matter, it plays a significant role in the rest of the enterprise categories. Participants who are involved in community advocacy groups, those who share/transfer information from the training, participants who are migrants, as well as those who practice savings, irrespective of gender decreases the probability of being unemployed. However only savings positively influence engagement in all the enterprises in the pooled data. For cocoa enterprise engagement, only those involved in advocacy groups and information transfer had a higher chance of engaging in it. Age, marriage, information transfer, savings, additional training, advocacy group influences cocoa farmers to diversify into non-cocoa business activities.

This study has key implications for policymakers, donors, as well as development practitioners who have keen interest in enhancing the welfare of youth and gender in cocoa growing communities. Aside training and capacity building, other factors impeding the participation of youth and women in cocoa production as mentioned earlier, should be looked at. This should factor in the development of gender-sensitive cocoa and non-cocoa production policies targeting the youth as they are the majority of the labor force and future duty bearers in the national economy. Now that there are numerous ways by which savings can be done, government in collaboration with the banking sector and telecommunication companies can promote the habit of saving among the youth. The amount saved can be beefed up with the institution of low interest loans for start-ups. Even though the area considered are cocoa growing communities, policies on employment should not focus so much on cocoa production to the neglect of the non-cocoa as a significant number of cocoa farmers diversify into non-cocoa enterprises.

Declarations

Author contribution statement

Abdul-Basit Tampuli Abukari: Analyzed and interpreted the data; Wrote the paper.

Abraham Zakaria: Performed the experiments; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Shaibu Baanni Azumah: Conceived and designed the experiments; Contributed reagents, materials, analysis tools or data.

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

The authors do not have permission to share data.

Declaration of interests statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.

Appendix 1

Unemployment rates (%) for various years among gender and age groups.

GLSS SURVEYS	Gender	15–24	25–44	45–64	Total
GLSS 1 (1987–1988)	Male	3.2	2.1	1.4	2.2
	Female	1.8	1.1	0.5	1.1
	Total	2.4	1.6	0.8	1.6
GLSS 2 (1988–1989)	Male	1.3	1.2	0.3	1
	Female	1.4	0.6	0.2	0.7
	Total	1.4	0.9	0.3	0.9
GLSS 3 (1991–1992)	Male	7.6	3.3	2.2	3.7
	Female	9.4	5.1	2.9	5.4
	Total	8.6	4.3	2.6	4.7
GLSS 4 (1998–1999)	Male	12.7	7.3	4.8	7.5
	Female	18.7	7.5	4.5	8.7
	Total	15.9	7.4	4.7	8.2
GLSS 5 (2005–2006)	Male	4.1	4	1.8	3.5
	Female	4.1	4	2	3.6
	Total	4.1	4	1.9	3.6
GLSS 6 (2012–2013)	Male	10.2	3.3	2.8	4.8
	Female	11.7	4.1	3.2	5.5
	Total	10.9	3.8	3	5.2
GLSS 7 (2016–2017)	Male	17.7	5.1	3.3	8.7
	Female	19	8.1	4.6	10.6
	Total	18.3	6.7	4	9.6

Source: GSS, 1989, GSS, 1995, GSS, 1996, GSS, 2000, GSS, 2008, GSS, 2014, GSS, 2019.