Determinants of Multiple Agricultural Technology Adoption: Evidence from Rural Amhara Region, Ethiopia

Mesele Belay Zegeye is a lecturer at Debre Berhan University, Ethiopia. His research interest includes Effect, Impact and effectiveness of policies, poverty, food security, technology adoption, and agricultural innovations and extensions, and agricultural productivity. Abebaw Hailu is a lecturer at Debre Berhan university, Ethiopia. His current research interests are development issue like urban and rural development, microeconomic and macroeconomics. In our country Ethiopia, where agriculture is at the heart of the economy. This study tried to examine the determinants of multiple agricultural technology adoption so as to raise the productivity of the agricultural sector. The result reveals that educational level, family size, off-farm participation, livestock, extension contact, credit access, advisory service, and farmers closer to plot, all-weather road, zonal town, and farmers with lower remittance income are the determinants of multiple agricultural technology adoption in Amhara Region. Abstract


INTRODUCTION
Agricultural or rural developments are a central issue to improve the welfare of rural households such as poverty, food insecurity and environmental sustainability. The various manifestations of poverty are found disproportionately in rural areas: low income, vulnerability to shocks, a poor infrastructure facility, political marginalization, and exposure to the degradation of natural resources, and make their livings from smallholder agriculture. This implicates the need for improving agricultural productivity at farm level, and thereby improving farmer's welfare (McCalla, 2001;and Admassie & Ayele, 2010). Furthermore, agriculture, with its high contribution to country's GDP, exports, and employment, it is an essential motor of growth in most developing countries.
According to Wik et al. (2008), growth originating in agriculture is about four times more effective in reducing poverty than other sectors. For this reason, policies that increase agricultural productivity can have a significant impact on poverty reduction. This is possible if and only if modern agricultural technologies are properly transferred and diffused so as to increase productivity. The green revolution introduced high yield varieties, fertilizers, pests, and others in developing countries, but the take up of these technologies in many developing countries has been uneven and have a low rate. In many areas traditional farming practices still dominate and the take up of the new technologies remains limited (De Janvry et al, 2017). This is true for SSA countries where the agricultural sector dominates and is characterized by low productivity due to the low rate of technology adoption (Asfaw et al, 2012).
In eastern African countries specifically, Ethiopia's agricultural and rural development policy has been aiming at enabling efficient use of modern agricultural inputs and practices (land management, fertilizer, chemicals, soil and water conservation, improved seed, advisory and credit service) among smallholder farmers for increased productivity by research and extension system (Tefera et al., 2016). Since, land is scarce; the feasible way to improve agricultural productivity in the country is through agricultural intensification (investing in new technologies) (Mohammed, 2014). However, the uptake of new technologies is low despite the fact that adoption has a direct impact on increasing yield and income generation as well as nutrition level (Woldegiorgis, 2015). Hence, their production and productivity are remaining very low; but the rapid population growth together with limited farm size and low productivity threatens the lives of the population for the future.

P a g e 3 | 16
There are different scholars have been conducted on determinants of agricultural technology adoption in rural parts of Ethiopia. To mention some of them are CIMMYT (1993); Admassie and Ayele (2010); Adekambi et al., (2009);Challa and Tilahun (2014); Asfaw et al.(2012) stated that many factors can be associated with a lower rate of agricultural technology adoption: manly influenced by factors of socioeconomic, institutional, access and distance related factors, and plot characteristic. Specifically: credit constraint, lack of insurance, high transaction cost, behavioral inadequacies and information about new technologies (extension services, agrodealers, farmers' cooperatives, and social networks), farm size, number of family size, age and sex of farmers, perception of farmers, off-farm income, distance to urban centers and development agents, plot distance, soil quality and cost of inputs. The study conducted by Sebsibie et al (2015) conducted on agricultural technology adoption and rural poverty in Amhara Region, Ethiopia (only on fertilizer use); (Amare, 2018): determinants of adoption of wheat row planting in Wogera district of Amhara region; Wudu (2017): determinants of adoption of improved wheat technology in Gozzamen district of Amhara region; Gebru (2016), the determinants of modern agricultural inputs adoption and their productivity in Amhara and Tigray regions on single adoption. The previous study focused on the determinants of single agricultural technology adoption in Amhara Region, Ethiopia. Hence, none of them examined farmers' decision to adopt multiple agricultural technologies Amhara Region, Ethiopia. Furthermore, some of the studies have also a small area coverage and small sample which might not be helpful to reach on a general conclusion at regional level. However, given the major research and knowledge gap, this study intends to examine the determinants of multiple agricultural technology adoption in Amhara Region, Ethiopia. Furthermore, this study includes remittance income as an additional factor of adoption.

Data Description
The study is conducted in the rural part of Amhara region of Ethiopia. The data for this study were obtained from the Ethiopian Socio-economic Survey (ESS) during 2015/16. The survey covers all the regions of Ethiopia; however, the data is argued to be representative for regional estimation in the most populous regions of Tigray, Amhara, Oromia, and Southern Nation and Nationality people's regional state. The survey covered a wide range of topic such that, covers a range of topics including demography, education, health, labor, welfare, agriculture, food security, and shocks. The survey meets Ethiopia 's data demand gaps and is believed to be of high quality, and accessible to the public. The survey covers rural areas, small towns, medium and large towns in all regions except some exceptional zones of Afar and Somali region.
Accordingly, the study considered 656 farm households drawn from rural Amhara regional state as a sample of the study.

Multinomial Adoption Selection Model
The choice of agricultural technologies is made according to the expected benefits from the adoptions of specific choices given their limitations. The common starting point is a random utility framework, in which the utility of each alternative is a linear function of observed characteristics plus an additive error term. Economic theory dictates that farmers adopt a single or a combination of technologies that can maximize their utility. This implies adoption occurs if the utility of the chosen package is higher than the utility of the other alternatives. However, the utility that gained from adopting agricultural technology is not observed but only its choice of technology, one can assume a random utility model which states conditional probability choice given farmers choice. To formalize this, consider the following latent variable: * = + η …………………………………………………………………. (1) Where: Aij* is a latent variable which describes the i th farmer's behavior in adopting the alternative package of technology J (j = 1, 2, …….m) with respect to another alternatives K.
Z's are a vector of observed independent variables (household characteristics, farm-level factors, institutional factors, biophysical factors and technology aspects) and η are unobserved characteristics which are relevant to the farm household's decision maker but are unknown to the researcher.
The farm household i will choose a package of j-technologies with respect to adopting any other technologies of k if it provides greater expected utility than any other alternative k, k ≠ j, It is assumed that the covariate vector of Zi is uncorrelated with the unobserved error termη , i.e., E (η |z ) = 0. Assuming that η are independent and Gumbel (identically) distributed (independence across utility functions and identical variance), that is under the independence P a g e 5 | 16 of irrelevant alternatives (IIA) hypothesis; this model leads to the selection of a multinomial logit model.

Multinomial Logit Model
Under adoption of multiple agricultural technologies, the number of alternatives that can be chosen is more than two; we can apply the multinomial discrete choice model to estimate simultaneously the effects of the explanatory variable on the adoption of different agricultural technologies. The variable Aij is a multiple choice variable and can be consistently estimated using a limited dependent variable (Maddala, 1983). In this regard, this study applied the multinomial logit model over other since the model is simple to calculate the choice probability and computers can maximize the resulting likelihood function even for a large number of choices. And also the result obtained from the model is more stable than others when independence of irrelevant alternatives(IIA) assumption fulfilled. (Kropko, 2008) also shows the multinomial logit model nearly always provides more accurate and realistic results than others even if the independence of irrelevant alternatives (IIA) assumption is severely violated.
The probability of choosing alternative packages of J using multinomial logit model (Pij) can be computed as: Where, Pit is adoption of multiple agricultural technologies, αi is the vector of parameters and Xi is a vector of all explanatory variables those are age of the household head, sex of household head, number of family size, education of the household head, total land size, distance to market, distance from zonal town, distance from the all-weather road, total livestock, credit access, extension services, getting advice, amount of farm remittance income, off-farm activities, far from homestead, plot potential wetness index, soil fertility quality and ownership of plots of land. The interpretation of the multinomial logit model is relative to the reference or base category group is difficult, even if this study used non-adopters as a base category. The coefficients need to be adjusted to be marginal effects in the case of the logit model.

Modeling Multiple Technology Adoption
The adoption of multiple agricultural technologies can be modeled in the setting of a framework. The adoption variable for this study is generated from the combinations of the organic fertilizer, inorganic fertilizer and herbicide technologies. The estimations of the adoption of alternative agricultural technology packages are estimated using multinomial logit model.
Note: Each element in the combination is a binary variable and for chemical fertilizer (F), organic manure (M) and herbicide adoption (H), and the subscripts represent 1 = adoption and 0 = nonadoption.  The descriptive statistics of explanatory variables for the eight combined alternative packages considered in this study are presented in Table 4. The explanatory variables' mean value of non-adopters (F0M0H0) is used as a base category to compare with mean values of alternative adopters (F1M0H0, F0M1H0, F0M0H1, F1M1H0, F1M0H1, F0M1H1 and F1M1H1). The result shows that the mean comparison test between adopters and non-adopters are significantly larger for adopters and the values are different across the different alternatives. For instance, a set of household characteristics like sex (adopters are more of male headed), regarding education, adopters have high education level. Moreover, on average adopters have large family size than the non-adopters. On the other way, the mean age of the household head for adopters is lower than the non-adopters.

2.3.Measurement and description of variables
Total farm size, remittance income, livestock wealth measured in TLU, and off-farm employment are used as a means to describe the economic status of the household. Average farm size for non-adopter is lower than adopters (except for herbicide adopters). The mean livestock wealth and the proportion of households with off-farm activity participation for adopters are higher than non-adopters. On average, remittance income of the household is larger for non-adopters (F0M0H0).
The mean distance from road, market and zonal town are larger for non-adopters. But adopters of manure, herbicide and a combination of the two packages, are higher than non-adopters indicating that as the distance increases farmers adopt the kind of technology that are easy to carry and nearby available technologies so as to reduce transportation costs.
Institutional factors such as extension contact, advisory service, credit access, and tenure security are other factors which affect the adoption decision. The first two are related with farmers' access to information on different packages and its profitability while access to credit indicates farmers' ability to finance their purchase of modern technology under cash constraints. The result shows these supports are higher for adopters. This may indicate that the institutional support system has long been a major factor for modern agricultural technology adoption even if its support has remained low (Kassie et al, (2010). Furthermore, adopters have significantly higher mean values in terms of microclimate indicators like plot potential wetness index, and lower mean values in terms of plot distance from homestead. Similarly, the proportion of non-fertile soil plot is significantly higher for non-adopters than adopters.

Econometric analysis
In econometric analysis, the study applies a method of analysis of maximum likelihood estimation technique for the purpose of estimating the multinomial logit functions. The model fits the data reasonably well. Various post estimation tests were made to check the validity of the model. These are Wald test is used to ensure that all regression coefficients are jointly equal to zero is rejected with [ 2 (133)=( 478.82); P0.000], The Hausman test result for test of independence of irrelevant alternatives assumption, Variance inflation factor for continuous and correlation matrix for categorical variable also have been seriously conducted for multinomial logit model. The results confirm that there is no serious multicollinearity problem across the explanatory variables. And finally, robust regression has used to control the problem of hetrosckedasticity and non-normality.   Kassie et al. (2010). This is because farmers with more educational years are more likely to adopt as they able to acquire, analyze and evaluate information on modern technology, market opportunity and its implied benefit.

Multinomial logit estimation results
The family size of farmers has a positive and significant effect on the adoption decision of manure and a combination of manure with herbicide packages. This implies the more family member the more the adoption will be, because adoption of multiple farm technologies requires and attracts more labor force for agricultural activities Kassie et al, (2010). Livestock wealth measured by TLU significantly and positively affects the adoption decision of multiple technologies. This is because of farmers who possess a flock of livestock are more likely to adopt than the have-not as it helps to get improved technology (as income means and source of manure input), consistent with Mulugeta & Hundie (2012). Farm households who participate in off-farm activity are also more likely to adopt chemical fertilizer and herbicide packages. This is because of participating in off-farm activities can generate income and solve the problem that the farm household's face while intending to purchase farm technologies, consistent with Kassie et al. (2010); and Mulugeta & Hundie (2012).
The other important factor is farm remittance income, and the result shows that the impact of remittance income on the adoption decision is negative and significant. The descriptive analysis also supports this finding. From the descriptive analysis it was found that the remittance income is higher for non-adopter. One plausible reason for this is the household's way of spending the money from remittance. Most often income from remittance will be used for daily consumption purpose than investing in agricultural development. This is supplemented by (Tuladhar et al., 2014). The higher remittance inflows to households, and subsequently, the higher income buffer, might have increased the opportunity cost of engaging in agriculture, resulting in reliance on remittance income more than the income from the agriculture sector.
Distance to market, road and zonal town are significant proxy variables which capture the relationship between access to market information, access to input and farm households' technology adoption. Distance to zonal town has a strong and negative effect on the adoption of package F1M0H0, F1M1H0, F1M0H1, and F1M1H1. Distance to all weather roads has also negative effect on the adoption of package F1M1H1. This result may verify that those framers who live away from service centers such as urban centers, development agent, and market place are less likely to adopt farm technologies because of farmers could have less access to information on improved technologies and high production cost, and hence are unlikely to adopt new or improved technology; similar with Admassie & Ayele (2010);and Hailu et al. (2014). However, distance to road and zonal town affects positively the adoption of package F0M1H0 and F0M0H1 respectively. This is because of as the distance to road increases; farmers choose nearby available and easy to carry type of technology packages so as to reduce production costs.
Institutional support factors such extension contact, advisory service, credit access and tenure security are considered equally important to understand farm household's technology adoption decision. That is, access to extension visit and advisory service has a positive and significant relation on the adoption of alternative technologies. This is because the extension contact helps the farmers to raise their awareness about the characterization and attributes of the technology, use and their impact. Extension gives detailed information, training and advisory services about the source, use and importance of the technologies to the farmers and engaging in input distribution. Advisory service on the other hand indicates that having access to regular and frequent advisory services by development agents, farm cooperatives and meetings plays a fundamental role in the dissemination and adoption of farm technology. Access to credit has a positive impact on the adoption decision of different packages. This is because of credit access solves income problems that household could face while they want to purchase agricultural technologies; and hence paves the way for timely application of modern farm inputs.
The positive and significant effect of tenure security for a combination of chemical and manure fertilizer shows that when farmers are secured for their own land, the more likely in the adoption practice since they can make long-term investment, similar with Admassie & Ayele (2010); Tefera et al. (2016); Sebsibie et al. (2015); (Mohammed, 2014); and Hailu et al. (2014).
Plot characteristics and microclimate variables are another set of factors considered in explaining household's likelihood of technology adoption, such as soil quality, plot wetness and plot distance. Plot distance from the homestead is negatively affects the adoption of package F0M1H0, and F1M1H0. This is because of as plot is far away from the homestead, the less likely will be on time plot preparation, weed, harvest and input utilization and hence farm households are less likely to adopt, confirmed with Kassie et al. (2010); and Hailu et al. (2014).
On the other way also, plot potential wetness index has a positive impact on the probability of the adoption of full packages (F1M1H1). This is because of as the wetness of the plot increases (maintains vascular plant species richness, soil pH, groundwater level and soil moisture) the more likely the adoption of the agricultural technology. This is in line with the finding of Sebsibie et al (2015). Most of the explanatory variables like sex of the family head distance to market, farm size and soli quality, even if they are not statistically significant, they have the expected signs with the adoption decision.

Conclusion and Policy Recommendation
Agriculture is the mainstay of Ethiopian economy, and where severe poverty is the main challenge, reducing poverty is the preliminary concern, by boosting production and productivity of agriculture through the use of improved agricultural technology is considered as a major solution. Thus, the main objective of this study was to explain factors affecting the adoption of multiple farm technologies in rural Amhara region. Multinomial logistic regression model was employed to identify the determinants of adoption of multiple farm technologies in rural Amhara Region, Ethiopia. The result of the study revealed that farm household's decision to adopt improved or new farm technologies is manly influenced by factors of socioeconomic, institutional, access to information and distance factors, and plot characteristic. More specifically, the adoption of multiple agricultural technologies is positively and significantly affected by education level of the head, household size, off-farm activity, livestock wealth, farmers contact with extension and advisory service, farmers having credit access, farmers secured for their own land, and plot potential wetness index. On the other hand, the adoption of farm technologies is negatively and significantly affected by age of the family head, distance from zonal town, road and plot distance from the homestead, and farmers getting remittance income. However, the effect of distance from road and zonal town for the adoption of herbicide and manure technologies is positive, indicating as the farm lives far away from urban centers and development agents, farmers adopt nearby available technologies and easy to carry type technologies so as to reduce their cost of production.
Thus, in terms of policy implication, polices for strengthening the access to information about the farm technologies by increasing the availability and quality of extension service, encouraging the participation of farmers in training centers and providing advisory service to increase the adoption by farm households. The influence of extension service, trainings and advisory services can counter balance the negative effect of lack of years of formal education on the overall decision to adopt farm technologies.
Distance from zonal town and road matters the adoption decision of multiple farm technologies. Thus, the government and other stakeholders should support the access of these infrastructural services.
Access to credit is also found a significant factor that positively affects technology adoption.
This may call for the need and expansion of credit institutions in rural areas where financial constraint is a major challenge for farmers while in adopting farm technologies.
Lastly, the effect of remittance income on technology adoption is negative. So, Policy measures aimed at channeling remittances to investments in productivity-enhancing agricultural capital and inputs might help increase agriculture yield.
Overall, better targeting of agricultural technology adoption and appropriate rural institutions on rural farmers might be the main vehicle for increasing farm technology adoption by the rural poor.