Six indices of social capital among the cocoa-based farming households are identified.The density of membership to associations is 0.636, which means that cocoa-based farming households belong to 6 out of 10 associations.Households belong to various associations in order to promote and protect their business interests.The degree of heterogeneity is considerably high in the area of study.This suggests heterogeneous characteristics such as different ethnicities, occupations, religion and neighbourhoods among the cocoa-based farming households.The decision-making value is fairly high in the associations, which implies that most of the cocoa farming household members are actively involved in decision making within the social group they belong.The meeting attendance value is 0.318, which means that the cocoa-based farming households do not attend most of the scheduled statutory meetings.This could be attributed to the fact that the households trekked an average of 0.70 km to designated meeting points.Hence, they only attend meetings whenever important decisions are to be made.The cash contribution value is 0.652, which means that the cash commitment to associations by cocoa-based farming households is relatively high.This shows that cocoa-based farming households are committed to contributing cash to their respective associations.Also, labour contribution has a value of 0.541, which means that the mean labour contribution is 54 days annually.This result compares favourably with Ajani and Tijani.The aggregate social capital, which is the multiplicative value of density of membership, heterogeneity index and decision-making index is 0.568.The result shows that a fairly high level of social capital exists among cocoa farming households in the study area.The factors that influenced the decision to participate in social groups are shown in Table 3.These characteristics include asset, age, years of education, gender, farm size,ebb and flow trays land tenure, loan interest rate, and extension visit.

The negative signs of marginal effects reduce probability of household participation in social groups while positive signs increase the probability of participation.Assets of households significantly affects the probability of participating in social groups.An additional unit of households’ asset decreased the decision to participate in social groups by 10.9 percent.This implies that increase in asset ownership decreases probability to participate in social groups.The age of household head significantly affected the probability of participating in social groups.A year increase in age of household head increased the decision to participate in social groups by 54.8 percent.This could be attributed to the fact that social groups might prefer older farmers to their younger counterparts, because older farmers are more responsible and secured to participate.In addition, due to some cultural beliefs in Africa, younger people might be prevented from participating in social groups.Years of education of household head significantly affected the probability of participating in social groups.An increase in years of education of household head increased the decision to participate in social groups by 7.7 percent.This implies that education gives farmers the ability to access and comprehend information regarding the terms and conditions required to participate in social groups.The gender of household head significantly affected the probability of participating in social groups.A male household head increased the decision to participate in social groups by 17 percent.This could be attributed to the fact that male headed families are willing to take more risk than female headed families.In addition, due to some social-cultural values and norms of Africans, male farmers have more freedom to participate in different social groups compared to the female farmers.The farm size of household significantly affected the probability of participating in social groups.An increase in hectares of farm size increased the decision to participate in social groups by 28 percent.The result implies that farmers with large farms possess the ability and collateral to participate in social groups.Land tenure status of household head significantly affected the probability of participating in social groups.This could serve as a push factor to participate in social groups in order to put resources to optimum use.

Loan interest rate significantly affected the probability of participating in social groups.A one percent increase in interest rate decreased the decision to participate in social groups by 46.6 percent.This implies that high interest rate constitutes a hindrance to loan access.Extension contacts significantly affected the probability of participating in social groups.A contact of household with extension agents increased the decision to borrow by 5.4 percent.This is because extension services could provide farmers with essential information regarding participation in social groups.The results of the factors influencing cocoa-based farming households’ level of participation in social capital groups are presented in Table 3.These include age of household head, years of education, membership in agricultural organization, off farm income, land tenure, interest rate, distance to credit sources, extension visits, decision making, cash contribution, and labour contribution.The age of household head significantly affected level of participation in social groups.A year increase in age of household head increased the level of participation in social groups by 3.164 units.This could be ascribed to the fact that as farmers get older, they become more productive and increase their level of participation in social groups.Years of education of household head significantly affected level of participation in social groups.An increase in years of education of household head increased the level of participation in social groups by 0.662 units.This is because education equips farmers to make informed decisions about their level of participation in social groups.Membership in agricultural organisations significantly affected level of participation in social groups.Being a member of agricultural organisation increased the level of participation in social groups by 2.085 units.This is attributed to the fact that membership in an agricultural organisation increases access to credit.Non-farm income significantly affected level of participation in social groups.An increase in non-farm income of households decreased the level of participation in social groups by 0.041 units.This is because farmers that earn substantial amounts of non-farm income would less likely need external funds.Land tenure significantly affected level of participation in social groups.Households with secure land increased the level of participation in social groups by 1.858 units.This could serve as a push factor to farmers to increase their level of participation in social groups.

Loan interest rate significantly affected level of participation in social groups.A one percent increase in loan interest rate decreased the level of participation in social groups by 1.225 units.High interest rates constitute a hindrance to the level of participation in social groups.Distance to credit source significantly affected level of participation in social groups.A kilometre increase in the distance to credit source decreased the level of participation in social groups by 0.010 units.Farmers who live near the social group’s designated building have a location advantage, which increases their level of participation in social groups.Extension visits significantly affected level of participation in social groups.Households’ contact with extension agents increased the level of participation in social groups by 0.768 units.Extension services provide essential information to farmers regarding participation in social groups.Decision-making significantly affected level of participation in social groups.An increased unit of the decision-making index increased the level of participation in social groups by 1.424 units.This is ascribed to the fact that decision making keeps individuals abreast of the association’s benefits.Cash contribution significantly affected level of participation in social groups.A naira increase in cash contribution increased the level of participation in social groups by 9.923 units.This implies cocoa-based farming households who made adequate financial contributions to social groups have access to substantial amounts of credit compared to households who did not.Labour contribution significantly affected level of participation in social groups.An increase in labour contribution increased the level of participation in social groups by 3.353 units.This implies that cocoa-based farming households who make adequate labour contributions in social groups have access to substantial amounts of credit compared to households who do not.To test for validity of the instrumental variables used in the 3SLS estimation procedure, a correlation analysis between aggregate social capital, farm productivity and food security with the proposed instruments was carried out.The proposed instruments were length of residency, charity donation, membership in religious group,4×8 flood tray and membership in ethnic groups.The results of the correlation analysis are presented in Table 4.The membership in ethnic groups has significant correlations with aggregate social capital, but an insignificant correlation with farm productivity and food security.It also has the highest correlation coefficient with the social capital.This conforms to the findings of Adepoju and Oni.

The basic model is shown in the first column of Table 6.The rationale behind this model is to examine the farm productivity of the households while they are not involved in social capital activities.The Chi2 showed that the econometric modelling is appropriate and correctly specified.Age of household head significantly influenced farm productivity of the cocoabased farming households.This implies that a unit increase in age of household head decreased households’ farm productivity by 0.836 kg/₦.This is attributed to the fact that ability to do farm work and farm output reduces with ageing.Household size significantly influenced farm productivity of the cocoa-based farming households.This implies that a unit increase in household size increased households’ farm productivity by 0.865 kg/₦.This is attributed to the fact that family labour available for farming could increase farm output.The results agree with Atagher.Primary and secondary education significantly influenced households’ farm productivity.The implication of this is that a unit increase in primary and secondary education increased households’ farm productivity by 0.014 and 0.113 kg/₦, respectively.This could be traced to the fact that education empowers farmers to access required skills and to utilise existing resources on the farm to boost their productivity.Farm size significantly influenced farm productivity of the cocoa-based farming households.This implies that a unit increase in the farm size increased households’ farm productivity by 0.214 kg/₦.This is ascribed to the fact that resources on large farms would increase farm productivity.Interest rate significantly influenced farm productivity of the cocoa-based farming households.This implies that a percentage increase in interest rate decreased households’ farm productivity by 0.346 kg/₦.This is because high interest rate discourages farmers from applying for loans and the amount of the loans farmers receive.Correspondingly, this reduces the quantity and quality of farm inputs that the farmer buys and negatively affects productivity farmer.Loan time lag significantly influenced farm productivity of the cocoa-based farming households.This implies that a unit increase in loan time lag decreased households’ farm productivity by 0.660 kg/₦.This is because long loan time lags would delay the procurement of a loan, which implies that farm inputs will not be available to the cocoa farmers at the right time, quantity and quality.This affects productivity of the farmers due to the seasonal nature of agriculture.This model suggests that households’ social-economic characteristics, farm specific and credit variables play a significant role in improving farm productivity.The second column of Table 6 shows the inclusion of six additive forms of social capital variables identified in this study.These include density of membership, decision making, cash contribution, labour contribution, meeting attendance and heterogeneity.The rationale behind the model is to examine the farm productivity of the households while they are involved in social capital activities.This new model has a better farm productivity level as reflected in a Chi2 of 45.34.This suggests that households’ farm productivity improve as members become involved in the affairs of their social groups.This model shows that the effect of social capital on farm productivity can be traced to meeting attendance, decision making, membership density, and cash contribution.This finding is line with the findings of Balogun et al..Meeting attendance significantly influenced farm productivity of the cocoa-based farming households in the study area.This implies that a unit increase in attendance of meetings increased households’ farm productivity by 6.959 kg/₦.This is because farmers who recurrently attended group meetings have access to resources and information to improve their productivity.Decision making significantly influenced farm productivity of the cocoa-based farming households.This implies that a unit increase in active participation in decision of the group decreased households’ farm productivity by 4.824kg/₦.This means that farmers’ involvement in association matters is of no benefit to their farm productivity.Density of membership significantly influenced farm productivity of the cocoa-based farming households.This implies that a unit increase in the number of groups to which a farmer belongs increases productivity by 0.450 kg/₦.This is attributed to the fact that farmers’ commitment in many social groups enhances their access to loans, which can be used to increase their productivity.