Determinants of Land Conservation Technologies Adoption among Arable Crop Farmers in Nigeria : A Multinomial Logit Approach

As parts of the efforts to reduce land degradation and hence improve farm productivity, farmers in the area were introduced to various land conservation technologies. The study was thus conducted to investigate the determinants of land conservation technologies in the area. Data collected through a multi-stage sampling procedure were analysed with the aid of descriptive statistics and multinomial logit model. Findings revealed significant difference between socio-economic characteristics of the respondents such as age (p≤0.01), household size (p≤0.01), farm size (p≤0.01), value of livestock (p≤0.01) and off-farm income (p≤0.01). Findings further revealed that level of education, extension contact and land ownership significantly influenced farmers adoption of mulching, cover cropping and tree planting respectively. On the other hand, while age of respondents reduced farmers’ adoption of mulching and tree planting, membership of association increased the adoption of cover cropping. Government through her agencies and other development organizations should therefore put in place policy framework that would educate the farmers through regular extension contact and also carry out a review of land ownership rights in the study area. This would enhance farmers’ adoption of conservation options and hence, sustainable production.


Introduction
Land degradation is a major factor militating against agricultural productivity in Africa.This poses a great threat to livelihood of farm families in Africa, Nigeria inclusive.Mugonola et al. (2013) revealed that restoring agricultural productivity requires the promotion and adoption of farm-level conservation technologies that are meant to reduce soil degradation.Various agricultural and non-agricultural processes and practices are known to cause different forms of soil degradation, among which are soil erosion, soil toxicity, soil pollution, poor land use regime such as bush burning (Mugonola et al., 2013;Iheke & Onyenorah, 2012).Soil erosion is a worldwide problem, particularly in the tropics where rainfall is high and intense.It has been recognized for a long time as a serious problem in Nigeria (Kabubo-Mariara et al., 2010).There are various conservation technologies which farmers can adopt in order to reduce the severity of poor productivity and poverty.Conservation techniques can be mainly runoff management techniques or fertility sustaining techniques.Kabubo-Mariara et al., (2010) stressed that land conservation entails fertility management which involves the use of such techniques as mulching, cover cropping, tree planting, among others.The adoption of these technologies which usually involves risk among the farmers may be influenced by a number of factors.Bekele and Mekonnen (2010) in their work in Ethiopia revealed that land and non-land factors affect farmers' decision at household plot-level to adopt conservation technologies.The land factors include farm size and tenure arrangement among others while the non-land factors are asset holdings (e.g.livestock) and income level.They added that institutional factors such as access to extension and, membership of association affect farmers' decision to adopt land conservation measures.
Land conservation technologies are known to play an important role in improving farm incomes.For this reason, substantial investments have been made in research to improve agricultural technologies in various parts of the world, from the development of new crop varieties to new practices of land management (Kabubo-Mariara, 2006;Sheferaw & Holden, 2001).
Despite the identified benefits of land conservation technologies on income, environment and the livelihood of the farmers and family, a substantial number of farmers in Nigeria do not adopt these technologies.In order to ensure conservation, sustainable utilization and management of agricultural lands, studies on determinants of adoption of conservation technologies with emphasis on socio-economic characteristics of respondents become priority.Therefore, it is important to empirically determine the specific effects these factors have in the willingness or otherwise of farmers to adopt land conservation technologies.This study thus addressed the following pertinent questions: what are the socioeconomic characteristics of the respondents?What are the land conservation techniques adopted by farmers in the study area?Which factors determine farmers' adoption of conservation technologies?The objective of the study was to investigate the factors influencing farmers' decisions to adopt land conservation techniques in the study area.

Theoretical Model
Land conservation option appears as a discrete choice set rather than a continuous one.When the dependent variable takes more than one value, the multinomial logit model is often appropriate (Akinola et al., 2011).The multinomial logit model was employed to package the various categories of land conservation practices into a three-model scenario.The model was employed instead of Tobit, logit or probit model because they assume that non-adopter of a given practice does not adopt any other as they only allow zero or one dependent variables.This is because when there is more than one practice to choose from, that the farmer does not pick one does not mean he is a non-adopter.Hence, non-adoption of one technology or practice does not necessarily puts the farmer in nonadopter category.This supports the model appropriateness for the various conservation options.

The model was specified as
Which implies that the utility, i U , of choosing a particular practice is a stochastic linear function of farm, farmers and practice specific attributes ( i X ).In this Multinomial logit, the probability, of choosing a given practice, j is equal to the probability that the utility of that particular technology is greater than or equal to the utilities of all other soil fertility technology in the model.The dependent variable in this model was a discrete variable taking the value 0, 1, 2, and 3, for cases of no-adoption, mulching adoption, cover cropping adoption and tree planting, respectively.

Study Site
The study was conducted in Ondo State, Nigeria.The state was purposively selected for the study owing to relative incidence of land degradation.It lies between Longitude 4 0 30 1 and 6 0 00 east of the Greenwich Meridian and Latitude 4 0 45 1 and 8 0 15 1 north of equator.The state is located on tropical coastal wetland with mean annual rainfall of about 2800mm, and mean number of rainy days of about 170.The mean relative humidity falls between 70-80%, mean annual temperature is about 27.8°C.The land area is about 14,798.8square kilometres with varying physical features like hills, lowland, rivers, creeks and water bodies.The predominant occupation in the area is farming which is characterised by smallholder farmers, cultivating both permanent and arable (such as yam, cassava, maize and cocoyam) crops for family consumption, market and cash.Farming activities are usually carried out using simple farm tools with limited application of modern implements.The total population and the population density of the state are 3,460,877 and 233.9, respectively (Ondo State Ministry of Economic Planning and Budget, 2010).

Sampling Procedure and Data
Multi-stage sampling technique was used in selecting respondents for the study.In the first stage, Ondo state was stratified into two agro-ecological zones based on the state's Agricultural Development Programme (ADP) classification.These are Ondo and Owo zones.Ondo zone consists of coastal forest and mangrove swamp forest while Owo zone consists of moist lowland forest, and forest savannah.The second stage involved the purposive selection of three local government areas (LGAs) from each of the zones based on the past records of land degradation.These are Irele, Odigbo and Okitipupa in the Ondo zone and Akoko South West, Ose and Owo in Owo zone.The third stage involved a random selection of 2 villages from each of the LGAs.In the final stage, 20 respondents per village were randomly selected making a total of 240 respondents.Primary data were used for the study and were collected using structured questionnaires.Information elicited from the respondents include their socio-economic and institutional characteristics.Data collected were analysed with the aid of Statistical Package for Social Sciences (SPSS) version 16.0 and LIMDEP version 7.0.

Empirical Model
Farmers' decision to adopt or not to adopt a technology is assumed to be the outcome of a complex set of factors related to the farmers' objectives and constraints.In other words, there are certain factors -social and institutional factors that affect the likelihood that farmers will adopt a technology.Thus if each farmer and each technology can be classified based on a core set of variables, then it is possible that the probability of a farmer adopting that technology could be estimated (Amare, 2012).Data collected from the survey were analysed using descriptive statistics and econometric model (Multinomial Logit).In this study, the dependent variable is the land conservation technologies adopted by the respondents.The estimated model is specified as follows: The multi-areas explanatory variables included in this study were those variables, which were believed to have influence on investment decisions on land conservation technologies.These included households, farm and institutional characteristics, tenure related factors, poverty related variables and market access factors.The rationale for the inclusion of these variables was based on a priori expectation of agricultural technology adoption.
Previous studies revealed that ages of individuals affect their mental attitude to a new technology and hence influence adoption in a number ways.Young farmers have been found to be more knowledgeable about new practices and may be more willing to take the risk and adopt or invest than the older farmers.This implies that farmer's age and technology adoption are inversely related (Dereje, 2006;Akinola et al., 2010).. Therefore, it is hypothesized that farmer's age and investment decisions are expected to relate negatively.Household size (HHSIZE) is the number of persons that live under the same roof.A large household size working on the farm reduces the farms' external labour requirements and is hence assumed to positively affect adoption of land conservation technologies (Bekele & Mekonnen, 2010;Akinola & Owombo, 2012).Farming experience (FARMEXPR) is measured in number of years, that is, the year a respondent started farming on his own.Experience of the farmer is likely to have a range of influences on adoption.Experience will improve the farmer's skill in production operations.Farmers with higher experience appear to have often full information and better knowledge and are able to evaluate the advantage of the technology (Godoy et al., 2001;Clay et al., 2002).Education (EDUCATN) measures the literacy level of individual farmer.The study hypothesized that it will positively influence adoption of investment in soil conserving technologies since it enhances the ability to obtain, process, use new information and thus increase the ability of farmers to use their resources efficiently.Several studies (Bamire et al., 2002;Dereje, 2006) indicated a positive relationship between education and technological adoption.Total farm size (FARMSIZ) holding may serve as a good proxy for wealth, status and income levels.This variable is likely to have a positive effect on adoption of most practices, meaning that the larger the farm size the greater the likelihood that a farmer will invest in land conservation technologies.This is possible as a farmer who has relatively large plot of land can rent out part of his land to earn income and run his production activity.Therefore, the size of the land will positively affect level of investment in conservation practices (Akinola et al., 2010(Akinola et al., , 2011)).Extension visits (EXTENSN) refer to the number of contacts farmer had with extension agent to take advice in last cropping season.Therefore, extension contact is expected to have a positive influence on farmer's investment decision on soil conservation technologies.It is believed that frequent contacts will enhance the exposure of farmers to improved production package (Owombo et al., 2011;Kidane, 2001).
Wealth related factors included were value of livestock (VALSTOK), Off-farm income (OFFINCOM) and credit access (CREDACCS).Measures of wealth comprising value of livestock, off-farm income and credit access are hypothesized to positively influence adoption of land conservation technologies.They are generally considered to be capital that could be used either in the production process or exchanged for cash or other productive assets.
VALSTOK increases the availability of capital which makes investment in land conservation technologies feasible.Off-farm income has the same effect on conservation adoption.Credit access takes cognizance of farmers' access to sources of credit to finance the expenses relating to adoption of innovation.Access boosts farmers' readiness to invest in technological innovations.It is hypothesized that the variable has a positive influence on the profitability of technology adoption (Oluoch-Kosura et al., 2001;Gebremedhin and Swinton, 2003;Hagos and Holden, 2006;Kipsat et al., 2007;Jungel et al., 2009;Akinola et al., 2012).It was measured as a dichotomous variable with access being one, and zero for no access.
Tenure related factors included in the study were land ownership (LNDOWSIP) and plot age (PLOTAGE).
LNDOWSIP measured as dummy (1, if owned and 0, otherwise) is a measure of the rights individual farm operator has over his or her plots.It is hypothesized that those who own plot have greater confidence in taking investment risks on farm (Bekele and Mekonnen, 2010).PLOTAGEE measures the length in years since land has been held.Some studies (Bekele and Mekonnen, 2010;Kabubo-Mariara, 2010) had established positive relationship between investment decision and plot age.This study hypothesized that if the farmers are confident of using a plot for at least five years, the greater the tendency that they would adopt conservation investment.
Membership of association such as cooperatives has been found to influence the interaction and exchange of ideas among farmers (Bamire et al., 2002, Akinola et al, 2010).Farmers who are not members of associations are expected to have lower probabilities of adoption and low level of adoption of conservation technologies.

Socio-economic characteristics of the respondents
Results of the survey revealed that there were variations in the demographic and socio-economic characteristics of the respondents' categories.The mean age of the non-adopters of any technology was 50.3 years which was higher than that of adopters.The mean age of the adopters ranged from 43.3 years to 46.1 years.F-test reveals a significant difference between the means of the respondents (p≤0.01).The average year of schooling was 2 years among the non-adopters which was the least among the respondents' categories.The years of education ranged from 4 to 13 years among the adopters though the mean age was 13 years which was the highest among the groups.The mean farm size was 6 which was the least among the respondents' categories.The house hold size was 8 each of non-adopters and adopters mulching respectively and 7 among the adopters of cover cropping.The mean farm size was 2.8 hectare and ranged from 3.2 hectare to 3.9 hectare among the adopters.It was highest (3.9 hectare) among the tree planting adopters.There was significant difference between the means of the respondents' farm size (p≤0.01).The value of livestock in naira among the non-adopters was N7,740 which is the least among the respondent categories.The value of livestock in naira ranged from 11,980 to 26,902 among the adopters.F-value showed significant difference between the means of the livestock income (p≤0.01).The number of extension contact was also least (3) among the non-adopters.The contact ranged from 7 to 13 among the adopters.The least number of contacts among the non-adopters could be traced to the farmers' membership of cooperative society.The farmers in the area are well experienced.The non-adopters have an average of 28 years of farming experience, which is the highest among the respondents' categories while it ranged from 17.7 years to 21.4 years among the adopters.The higher experience among the non-adopters could be traced to two factors, viz; age and the fact that most were born into farming.Farmers in the area engage in activities other than farming.The non-adopters recorded an average of N1, 445 as off-farm income in the area.The off-farm income in the area ranged from N9, 899 to N38,662 among the adopters.The off-farm income was highest among the adopters of tree planting.The higher off-farm income among the adopters might be the reason for the technology adoption as posited by Bamire et al., 2002;Akinola et al.2010;Owombo et al.2011 andAkinola andOwombo, (2012).F-value showed a significant difference between the means of the farmers' off-farm income (p≤0.01).
The results also revealed that each of the respondents' group has used the plots for over one decade.The plot age among the non-adopters was 22 years.The age ranged from 12 years to 17 years among the adopters.The higher plot age among the non-adopters could be traced to factors such as age and land inheritance among the rural households.Analyses further revealed that while only 1.3 percent of the respondents had access to credit, 14.1 percent of mulching adopters, 14.9 percent of cover cropping adopters and 21.2 percent of the tree planting adopters had access to credit, respectively.This could be traced to respondents' membership of association in the study area.Membership of association was least among the non-adopters as 28 percent of them belonged to one association or the other.Ownership of land was fairly reasonable in the area.While 54.6 percent of the respondents owned the land on which they operate, 56.6 percent of the mulching adopters, 59.4 percent of the cover cropping adopters and 73.4 percent of the tree planting adopters owned the plots on which they operate.The higher proportion of tree planting adopters who owned land could be the reason for tree planting adoption because of permanent investment nature of tree planting technology.

Multinomial Logit Model Results
The results of the determinants of land conservation technologies among the respondents are presented in Table 3.The results in the table showed that the Log likelihood function, Restricted Log likelihood function and Chi-squared values were -81.0811, -118.8223 and 92.0882, respectively.These values show the fitness of the entire model and are significant at 1 percent level of probability.The results in the table furthers revealed that the coefficients of level of education, number of extension contact and land ownership were positive and significantly influenced the farmers adoption of mulching, cover cropping and tree planting technologies, respectively.This implies that an increase in the years of education of respondents by 1 year would increase farmers of adoption mulching, cover cropping and tree planting technologies by 2.1 percent, 1.2 percent and 4 percent respectively.This is in agreement with previous studies such as Bamire et al. (2002) and Akinola et al. (2012).An increase in the number of extension contact by 1 would increase farmers' adoption of mulching, cover cropping and tree planting technologies by 0.5 percent, 5.4 percent and 3.2 percent, respectively.Also, increase in the ownership of land would increase respondents' adoption of the technologies.Household size was positive and significantly influenced mulching and tree planting technologies adoption in the study area.An increase in the household size by 1 would increase probability of adopting by mulching and tree planting technologies by 5.5 percent and 1.7 percent, respectively.This agrees with the a priori expectation of the study and in addition conforms with Akinola et al. (2010) and Akinola et al. (2012).The coefficient of farm size was positive and significantly influenced mulching technology adoption.An increase in the farm size by 1 hectare would increase the probability of adopting mulching by 7 percent.This conforms to the a priori expectation of the study and in agreement with Akinola et al. (2010).The coefficient of plot age was positive and significantly influenced tree planting technology adoption.This shows that an increase in the plot age by 1 year would increase probability of adoption by 21 percent.This is in agreement with the a priori expectation of the study and in addition conforms to Bekele and Mekonne (2010) that plot age influences conservation technologies with permanent investment like tree planting and stone terracing.On the other hand, while the coefficient of age of respondents was negative and significantly influenced the adoption of mulching and tree planting technologies, off-farm income was negative and significantly influenced tree planting.This implies that an increase in the age of the respondents by 1 year would decrease the probability of adoption by 2.3 percent and 5 percent, respectively.This is in agreement with the expectation of the study that age and technology adoption are inversely related and conforms to the previous studies such as (Dereje, 2006;Nkoya, et al., 1997;Akinola et al., 2010).The coefficient of age was negative and does not significantly influence cover cropping technology adoption.The sign is in agreement with the expectation of the study.An increase in the off-farm income by 1 naira would decrease the probability of adoption by 7.6 percent.This implies that the increase in the off-farm income was diverted to activities other than farming.This does not conform to the expectation of the study but in agreement with Bekele and Mekonnen (2010) that an increase in the income of farmers may be diverted to activities other than farming.Note: ***= significant at 1 percent, **=significant at 5 percent and *=significant at 10 percent ++= significantly influence the three technology.

Conclusion
Land degradation is a major factor causing poor yield among rural farmers which has resulted in poverty among the rural farmers.The study thus investigated the determinants of land conservation technologies adoption among arable crop farmers in the study area.Data which were from primary source were analysed with the aid of descriptive statistics and multinomial logit model.Results from analysis revealed that adoption of conservation technologies was influenced by farmers' level of education, extension contact and land ownership among others.Therefore, government through her agencies and others development organizations should encourage farmers on the need for conservation technologies' adoption and put in place policy thrust that would educate farmers, improve extension services as well as review ownership rights to land.

Table1
contact a respondent had OFFINCOM Income in naira earned from off-farm activities N CREDIT Access to credit measured by the farmer's access to source of credit such as co-operative society at a reasonable cost. 1 if there was access, 0 otherwise VALSTOCK Money value of livestock owned by respondents in naira N ASSOC Membership of association 1 if you belong to association, 0 otherwise LANDOWSP Farmers' ownership of land on which he or she operates.1 owned, 0 otherwise PLOTAGE Year since land has been in used by respondents Year ε i Error term N = Nigerian currency; N 1= $0.006 Figure 1 r percent of the respon mulching adoption ( Table3.Average partial estimates of multinomial logit model for land conservation technologies adoption