Smallholder market participation and its associated factors: Evidence from Ethiopian vegetable producers

Abstract

Vegetable production is mainly practiced by smallholder farmers in Ethiopia and it is their main source of income, employment opportunities and improving food security. Although farmers are producing more surplus vegetables, they are not much linked with markets and thus their opportunity to diversify their livelihoods from vegetable production is very much limited. Thus, this study was required to identify factors affecting market participation among smallholder vegetable farmers in southwest Ethiopia. To get the sample respondents two-stage sampling procedures were employed and finally, 240 vegetable producers were selected. Both qualitative and quantitative data were collected from primary and secondary sources. To achieve the stated objective probit model was used and its result indicates that age of household head, household family size, education level of household head, labour market, market information and distance from the market place were statistically influencing market participation among smallholder vegetable producers. To enable smallholder’s participation in vegetable marketing government interventions is needed to strengthening institutional service, communication and infrastructure facilities. This study may be valuable input for smallholder farmers, policymakers and other stakeholders in revealing the gab in the performance of the current vegetable production system to realize the national development policy.

Keywords:

• farmers
• market participation
• vegetable
• probit model analysis
• ethiopia

PUBLIC INTEREST STATEMENT

Ethiopia has a worthy potential in the production of vegetables and the aim of production is mainly for subsistence, with less market-oriented activities. Despite the production potential and importance of vegetables, the production-market linkage is very weak. However, vegetables are commodities which have a higher value at market turning by more on consumption purpose than commercializing, and this is due to demographic, socio-economic, institutions, and other related factors. Although Ethiopian farmers are producing more of surplus vegetables, they are not much linked with markets and thus why their opportunity to diversify their livelihoods from vegetable production is very much limited. Thus, the special focus should be taken into account for smallholder vegetable producers on production-market linkages and on the way they can be more benefited.

Competing Interests

The authors declare no competing interests.

1. Introduction

About 1.3 billion people globally do not have enough food to eat and Sub-Saharan Africa host 22.8% of the undernourished population (FAO, 2019). Being the dominant livelihoods in Africa, agricultural growth unable to feed the growing population (AU (African Union), 2013). Smallholder agriculture has long been the dominant economic activity in the sub-Saharan region, and it will remain enormously important for the foreseeable future (Gollin, 2014). However, the subsistence nature of the sector and low market integration remains a big challenge. The Ethiopian case is no exception. Agriculture is the mainstay of the Ethiopian economy and more than 85% of the rural populations are engaged in agriculture. The livelihood of the smallholder farmers is also determined by this sector. This sector also plays a substantial role in the life and livelihood of most Ethiopians. It accounts for over 40% of GDP, over 80% of employment and 90% of foreign exchange earnings (Demese et al., 2010; Diao, 2010). This indicates that agriculture is the basis for every economic activity of the country. Agriculture determines the economic, social, and political system of society in developing countries like Ethiopia (Leykun. & Jemma, 2014). Ethiopian smallholder farmers are dependent on the cultivation of cereals (Salami et al., 2010; CSA (Central Statistical Authority), 2011; Tafesse et al., 2011). However, the agricultural production system of the rural people is featured by limited access to land, poor access to inputs, an underdeveloped irrigation system, inadequate market orientation, inadequate infrastructures, poor technology, inadequate extension advisory services and low output (Tilaye, 2010). Besides, the majority of smallholder farmers in Ethiopia are subsistence-based farming system and the linkage between production and consumption decision is very low (Muller, 2014). Their participation in subsistence farming does not ensure their food security and household welfare.

The Ethiopian government has formulated a series of policies, strategies, and programs to promote agricultural development to ensure food security and build resilience. The government has also developed the second Growth and Transformation Plan for the period 2016–2020 to become a middle-income country by 2025 by improving agricultural productivity and its commercialization. Among the strategies, market-oriented agricultural production policies are the central one (Mekonnen, 2015; Shiferaw & Teklewold, 2007) and the government tries to promote production and marketing of high-value agricultural products to increase the competitiveness of farmers in national and international markets (Tufa. et al., 2013). However, smallholder farmers are unable to benefit from such policy interventions due to unimproved varieties, high transaction costs, lack of infrastructures and inadequate extension services (Gebremedhin and Hoekstra, 2007). Thus, commercializing subsistence farming is a very decisive and important pathway to ensure household food security and economic growth of the country (Abafita et al., 2016; Mitku, 2014). Commercialization also enhances the links between the input and output sides of agricultural markets and farmers’ participation (Jaleta et al., 2009). Leykun. and Jemma (2014) addressed that the average crop output and input market participation are 25% and 20%, respectively, in 2009. This indicates that market participation in rural areas is not above average. Even if the efforts made by the government to transform smallholder farmers from subsistence to commercial farming system, the performance has been considered below expectations (NPC (National Planning Commission), 2016). This poor performance is because of a lack of modern inputs and inefficient use of resources (Kindie, 2005) and following a traditional way of farming system, poor production technology, rain-fed dependent agriculture, and low output mode of production (FAO (Food and Agriculture Organization), 2011). In Ethiopia as well, studies also found that the commercial orientation of smallholder farmers for crop production is very low (ADAM, 2009; Adane, 2009; Bedaso et al., 2012).

Vegetable production is subsistence farming practiced by smallholder farmers in Ethiopia and its cultivation is considered as the supplementary to the production of main crops. Nowadays, these crops are the main sources of income for smallholder farmers and their demand is also growing in both national and international markets (Bezabih & Hadera, 2007; Yilma, 2009) and as the result, the participation of horticulture producers is increasing. Though farmers have an interest in participating in the production and marketing of horticultural products, their participation is very limited because of different factors especially for those farmers who are living in rural areas. Among these factors poor transport, inadequate infrastructure, high transaction costs, lack of market information, and lack of feasible partners (Abafita et al., 2016). Mitku (2014) argued that market participation of smallholder farmers is very limited and agricultural markets are also fragmented which increases the transaction costs and reduces farmer’s interest to produce products for the market. To tackle these problems increasing the participation of smallholder farmers in the marketing of horticultural crops is very crucial (Olwande et al., 2015). Commercialization of smallholder farmers is the way to bring their commodities to the market and becoming beneficiary as inclusive development (Arias et al., 2013). Gebreslassie et al. (2015) suggested that market-oriented patterns of crop production can be effective and productive through intensification and commercialization of agriculture.

In Ethiopia, vegetable production is not available in all parts of the country but southern parts particularly Jimma zone have a good potential in vegetable production which is mainly utilizing them for stable food subsistence, with less market-oriented activities. Despite the production potential and importance of vegetable crops, there has been limited study with regard to the commercialization of vegetables. However, vegetables are commodities which have a higher value at market turning by more on consumption purpose than commercializing, and this is due to lack of information and other related factors. Vegetable commercialization by smallholder farmers is determined by household characteristics, household resource endowments, institutional factors, infrastructural factors and market-related factors (Goitom, 2009; Adam et al., 2010). Although Ethiopian farmers are producing more of surplus vegetables, they are not much linked with markets and thus why their opportunity to diversify their livelihoods from vegetable production is very much limited. Thus, getting access to markets for vegetable marketing is important to diversify the livelihoods of smallholder farmers and reduces rural poverty (World, 2008). The reports from government offices indicate that there is little empirical evidence on smallholder vegetable commercialization and its associated factors in Ethiopia. Other studies carried out in Ethiopia focused on the commercialization of other horticultural crops. Moreover, those studies that worked out have been focusing on the proportion of output sold in the market. In this study, we address such research gaps in the study area.

Therefore, this research aims at linking smallholder vegetable producers with markets to enhance the demand of the products and increase the means of generating their income. Therefore, this study was conducted with the objective of examining factors influencing market participation among smallholder vegetable farmers and their associated factors in the study context. This may be valuable input for smallholder farmers, policymakers and other stakeholders in revealing the gab in the performance of the current vegetable production system to realize the national development policy.

2. Materials and methods

2.1. Description of the study area

The research survey was conducted at Jimma Zone by focusing on Dedo, Kersa and Seka Chokorssa Districts in southwest Ethiopia. The area was located at (Latitude 7.6667° or 7° 40ʹ north; Longitude 37° or 37° east; elevation 1744 m). The distance between the capital city of Ethiopia (Addis Ababa) and Jimma city is 353.83 km (220 miles). Jimma has a relatively cool tropical monsoon climate and it features a long annual season from March to October. Temperature in the study area is with daily mean staying between 20°C and 25°C year round (CSA, 2007)

2.2. Sampling methods and procedures

The study was conducted in the southwest part of Ethiopia by selecting sampled respondents based on determining factors and levels of accuracy required. In this regard, this survey was conducted in three districts (Kersa, Seka Chokorsa and Dedo) in the Jimma area and these districts were selected purposively on the basis of the better production potential of vegetables. Among the districts, eight kebeles were also selected purposefully where the production potential of vegetables is very high. Finally, 240 vegetable producer sampled households were selected using simple random sampling method assisted by probability proportion to size. Then, a household head for each sample households was interviewed by well-trained and qualified enumerators. Moreover, Table 1 shows the selection of final sampled households.

Table 1. The name of districts, kebeles and the final sampled respondents

Name of districts	Name of kebeles	Total pop. in each selected kebeles	Proportion of sampled households (%)	Total sampled households
Dedo	Waro kolobo	4322	12.08	29
	Ganjo Abbe	4026	11.25	27
	Ofole korti	4531	12.92	31
Kersa	Marawwa	3502	10.00	24
	Siba	4846	13.75	33
	Girma	4123	11.67	28
Seka Chokorsa	Shane kochi	5268	15.00	36
	Buyo kachama	4737	13.33	32
Total		35,355	100.00	240

Source: Own study (2017)

2.3. Vegetable production status of farmers

According to the survey made the type of vegetable produced in the study area includes: Cabbage, Potato, Tomato, Pepper, Onion, Carrot, Garlic, and Chili. However, Table (2) shows that the most common potential vegetable produced (Cabbage, Potato, Tomato Onion) in three districts particularly in eight kebeles as it is shown in Table (2). Based on these, 54.6%, 64.6%, 55.0% and 62.1% of sampled respondents were producers of Cabbage, Potato, Tomato and Onion, respectively. However, 45.4%, 35.4%, 45.0% and 37.9% were non-producers of Cabbage, Potato, Tomato and Onion, respectively.

2.4. Data types, sources, and methods of collection

To answer the research questions and objective of the study both qualitative and quantitative data were collected from primary and secondary sources. The primary data were collected from vegetable producers, agricultural experts, local leaders and other subject matter specialists on various aspects of commercializing vegetables. Primary data were collected from the respondents using pre-tested questionnaires, and structured interview schedule by well-trained enumerators closely supervised by the researchers. Moreover, restructuring has been done using a sufficient number of non-sampled respondents through pretest in order to suitably modify the questionnaires and facilitate smooth administration. Secondary data on previous research findings and reports were collected from kebeles, districts, agricultural offices, ministry of trade, trade and revenue offices.

2.5. Methods of data analysis

The unit of analysis in this study was households of vegetable producers. To analyze the collected data, descriptive statistics, inferential statistics and econometric model were used. The descriptive statistics (mean and percentage), inferential statistics (t-test and chi-square tests) were used. Our dependent variable is dichotomous, representing farmers’ decision to participate or not in vegetable marketing. For such a dummy dependent variable, binary logistic or probit model is appropriate (Gujarati, 1995). Both logistic and probit models yield similar results (Greene, 2002). Different authors used probit analysis on the decision to market participation (Egbetokun & Omonona, 2012; Hlomendlini, 2015; Mbitsemunda & Karangwa, 2017). Hence, to analyze the determinants of smallholder farmers’ participation in the vegetable marketing probit model was used. Participation in the marketing decision of the respondents was taken as the dependent variable with a value of 1 if the farmer participated and 0 otherwise. In this model, the probability that Y = 1 (the probability that the household participates in vegetable marketing) was estimated using the cumulative standard normal distribution function. The researchers opted to use the probit regression model to identify the factors that determine the decision of smallholders to participate in the vegetable market hence the dependent variable is dichotomous. Accordingly, the dependent variable assumes only two values; 1 if the household participates in the vegetable market and 0 if he/she does not. Assume that Y can be represented by market participation and the regression equation is representing market participation (dependent variable, Y) and we also have a vector of regressors X, which are assumed to influence the dependent variable (Y). Specifically, we assume that the model takes form;

(1) $\mathrm{P}\mathrm{r}\left(Y=1|X\right)=\left(X^{\prime }\beta \right)$(1)

Where, Pr denotes the probability of household’s vegetable market participation decision, and Φ is the Cumulative Distribution Function of the standard normal distribution. The parameter β is typically estimated by maximum likelihood. It is also possible to motivate the probit model as a latent variable model. Suppose there exists an auxiliary random variable;

(2) $Y^{\ast }=X^{\prime }\beta +\epsilon ,$(2)

where ε ~ N(0, 1). Then, Y can be viewed as an indicator for whether this latent variable is positive:

(3) $Y=1\left\{Y^{\ast }>0\right\}=\left\{\begin{array}{c}1if{Y}^{\ast }>0i.e.-\epsilon <X^{\prime }\beta \\ 0otherwise\end{array}\right.,$(3)

Therefore, Pr—is the probability at which an individual household participate in vegetable marketing in 2017 in the year represented by (MarkPart = 1)

β’sare the coefficients to be estimated

X1 Age of households

X2 Family size

X3 Educational level

X4 Labour availability

X5 Access to market information

X6 Access to transport services

X7 Demand for vegetables at market price

X8 Off-farm income

X9 Access to extension service

X10 Distance to market

X11 Perishability of vegetable products

From equation (1)(1) $\mathrm{P}\mathrm{r}\left(Y=1|X\right)=\left(X^{\prime }\beta \right)$(1)

(4) $\begin{array}{rl}& P\left(Y_i=1|x_i\right)=P(Y_i^{\ast }>0|x_i)=P\left(\beta x_i+u_i>0|x_i\right);\\ & u_i\sim N\left(0,{\sigma }^2\right)=P\left(u_i>-\beta x_i|x_i\right)=P\left(u_i\le \beta x_i|x_i\right)=\\ & \underset{-\mathrm{\infty }}{\overset{P{x}_i}{\int }}\mathrm{\varnothing }\left(\beta x_i\right)d{x}_i=dy/dxwhichisthemarginaleffects\end{array}$(4)

Where, $\mathrm{\varnothing }$ is the cumulative normal distribution function.

Hence, finally $P$($Y_i=1|x_i$)$=\mathrm{\varnothing }\left(\beta x_i\right)=\frac{1}{\sqrt{2{\sigma }^2\pi }}{e}^{-{(x_i-\beta )}^2}/2{\sigma }^2$

Pr(MarkPart = 1) = β0+ β1X1i+β2X2i+ β3X3i+ β4X4i+ β5X5i+ β6X6i+ β7X7i+ β8X8i+ β9X9i+ β10X10i+ β11X11i+ Ui

The following are a host of the explanatory variables that are potentially expected to explain the variation in the dependent variable, market participation.

2.5.1. Definition of variables and hypothesis

In the course of identifying factors influencing the participation decision in the marketing of vegetables, the main task is to analyze which factors influence the participation of vegetable products. Therefore, potential variables, which are hypothesized to influence market participation and vegetable products, are presented below. Market participation is dependent (dummy) variable that represents the market participation of the household that is regressed in the probit model. The dependent variable was vegetable commercialization (market participation). It was determined by different factors such as socio-economic, demographic and institutional factors. According to various literature demographic, socio-economic, and institutional factors influence market participation (Amao & Egbetokun, 2018; Giziew, 2013; Rabbi et al., 2017). For instance, younger, men, educated and lower family farmers tend to participate in vegetable markets (Giziew, 2013). Similarly, Rabbi et al. (2017) identified gender, age, number of a family member, household size, vocational training, and farm size as the major determinants of market participation, while Amao and Egbetokun (2018) report gender, farming experience, land ownership, farm size, and group marketing as the determinants of market participation. In the same manner, N. Kyaw et al. (2018) identified education status, total produce, access to roads and extension services improve the likelihood of market participation among rice farmers. The following explanatory variables were hypothesized to influence the market participation decision of vegetable farmers (Table 3):

Table 2. Vegetable production status of households

Vegetable producer status of households	Cabbage	Potato	Tomato	Onion
Producer	131(54.6)	155(64.6)	132(55.0)	149(62.1)
Non-producer	109(45.4)	85(35.4)	108(45.0)	91(37.9)
Total	240(100)	240(100)	240(100)	240(100)

Source: Own study (2017)

Table 3. Description of explanatory variables for probit estimation

Expanatory variables	Variable type	Measurements	Expected effect
Age of household heads	Continuous	Number of years	±
Family size of household heads	Continuous	Number of family member	±
Educational level of household heads	Continuous	Years of education	+
Labour availability	Continuous	Number of active family labour (age 15–65 years)	±
Access to market information	Dummy	1 = Yes, 0 otherwise	+
Access to transport services	Dummy	1 = Yes, 0 otherwise	+
Demand of vegetables at market price	Dummy	1 = yes, 0 otherwise	+
Off-farm income	Dummy	1 = Yes, 0 otherwise	+
Access to extension service	Dummy	1 = Yes, 0 otherwise	–
Distance to market	Continuous	In Kilometer	+
Perishability of vegetable products	Dummy	1 = Yes, 0 otherwise	–

Source: Own study (2017)

3. Results and discussion

3.1. Descriptive results

Table (4) indicates that the mean age of household heads was 35.9 years for market participants but 59.07 years was for non-market participants. This implies that young households were more participating in selling vegetables than old households and thus why the number of young households was very large in the marketing of vegetables. The t-test result shows that the mean age difference between the market participants and non-market participants was found to be statistically significant at the 1% probability level. Likewise, the mean family size was 4.33, while 8.49 was for non-market participants. This implies that the number of consumers was larger in non-participants than in market participants hence the size of the family does not go along with the consumption level of the households. The t-test result indicates that the mean difference between the family size of market participants and non-market participants was statistically significant at 1% probability level. Regarding the distance of farmers’ residence to the nearest market, it was found that the mean distance of the market participants was 3.87 km, while it was 7.82 km for non-market participants. This indicates that the majority of non-market participants were far from the market place as compared to market participants. The result of the t-test shows that the mean difference between the distance of household residence to the nearest market for the market participants and non-market participants was statistically significant at 1% probability level.

Similarly, we used the chi-square test to understand the associations between dummy predictor variables and dependent variables. The assumption of the chi-square test is that no relationship exists on the dummy variables in the population. In terms of labour availability, 41.80% of market participants had no enough labor but 58.20% of them had enough labor for marketing of vegetables. In the case of non-market participants, 69.49% of them had no enough labor but 30.51% of them had no labor problem. On another hand, market participants had more enough labour force than non-market participants in the marketing of vegetables and the chi-square test result shows that there is a statistically significant difference between market participants and non-market participants at 5% probability level. Another variable was access to market information, the (Table 5) revealed that 68.03% of market participants and 29.66% of non-market participants had access to market information but 31.97% of market participants and 70.34% of non-market participants had no access to market information for vegetable marketing. Similarly, majority of the respondents from both market participants and non-participants were unable to getting market information timely so that they were exposed to selling their vegetable products at low price at farm gate. According to the result of chi-square test, the difference between market participants and non-market participants was statistically significant at 5% probability level. Regarding access to transport service, 45.08% of market participants and 70.34% of non-market participants had no access to transport service but 54.92% of market participants and 29.66% of non-market participants had and the result of chi-square test reflects that the difference between market participants and non-market participants was statistically significant at 1% probability level.

Table 4. Summary statistics for continuous variables

Explanatory variables	Market participants		Non-market participants		t	P-value
	Mean	SD	Mean	SD
Age of household heads	35.9	12.76	59.07	12.28	14.32	0.0000***
Family size	4.33	2.41	8.49	1.99	14.55	0.0000***
Distance to Market place	3.87	4.01	7.82	4.37	7.31	0.0000***

Source: Own survey (2017); *** indicates significant at the probability level of 1%.

Table 5. Summary statistics for dummy variables N (%)

Explanatory variables	Response	Market participant	Non-market participant	Total	χ^2	µ
Labour availability	No	51(41.80)	82(69.49)	133(55.42)	18.61	0.000***
	Yes	71(58.20)	36(30.51)	107(44.58)
Access to market information	No	39(31.97)	83(70.34)	122(50.83)	35.34	0.000***
	Yes	83(68.03)	35(29.66)	118(49.17)
Demand of vegetables	No	66(54.10)	84(71.19)	150(62.50)	7.47	0.006***
	Yes	56(45.90)	34(28.81)	90(37.50)
Access to transport services	No	55(45.08)	83(70.34)	138(57.50)	15.66	0.000***
	Yes	67(54.92)	35(29.66)	102(42.50)
Participation in off-farm activities	No	43(35.25)	81(68.64)	124(51.67)	26.79	0.000***
	Yes	79(64.75)	37(31.36)	116(48.33)
Access to extension service	No	93(76.23)	78(66.10)	171(71.25)	3.01	0.083*
	Yes	29(23.77)	40(33.90)	69(28.75)
Perishability of vegetable	No	77(63.11)	85(72.03)	162(67.50)	2.18	0.140
	Yes	45(36.89)	33(27.97)	78(32.50)

Source: Own study (2017), *** and *, represents statistically significant at the probability level of 1% and 10%, respectively.

The case of participation in off-farm activities, 64.75% of market participants and 31.36% of non-market participants were taking part in off-farm activities but 35.25% market participants and 68.64% of non-market participants were not participating in it. This indicates that vegetable market participants were more participating in off-farm activities than non-market participants and the chi-square test result shows that the difference between market participants and non-market participants was statistically significant at 1% probability level. In terms of access to extension services, majority of the market participants (23.77%) were getting less access to extension services than non-market participants (33.90%), especially on market price, costs and time of selling. Moreover, from the total of sampled respondents, 28.75% had got access to extension services but 71.25% did not get it. This shows that there was the problem of inadequate extension services delivered to smallholder vegetable producers in the study area. The result of chi-square test indicates that the difference between market participants and non-market participants was statistically significant at 5% probability level regarding access to extension service. About 45.90% of market participants and 28.81% of non-market participants their vegetable products had good demand at market level but 54.10% of and 71.19% of vegetable producers for market participants and non-market participants had no demand, respectively, and the chi-square test result shows that the relationship between market participants and non-market participants was statistically significant at 1% probability level.

Finally, regarding storage facilities, among the market participants of vegetable producers, 63.11% of them were not getting access to storage facilities but 36.89% of them had. In the case of non-market participants, 72.03% of them had storage facilities but 27.97% had not. However, the chi-square test result shows that the difference between market participants and non-market participants was not statistically significant.

The results in (Table 6) indicate that 27.05%, 21.43% and 44.26% of the market participants were illiterate, completed primary and secondary education, respectively, whereas 66.95%, 29.66% and 3.39% of the non-market participants were illiterate, primary and secondary education, respectively. This indicates that the majority of the non-market participants (66.95%) were illiterate when compared to market participants (27.05%).

Table 6. Summary statistics for categorical variables

Explanatory variable	Response category	Market participant	Non-market participant	Total
Educational level of household heads	Illiterate	33(27.05)	79(66.95)	112(46.67)
	Primary	54(44.26)	35(29.66)	89(37.08)
	Secondary	35(28.69)	4(3.39)	39(16.25)

Source: Own study (2017), ***, represents statistically significant at the probability level of 1%.

3.2. Results of the econometric model

In this section, an econometric analysis was performed to identify factors that influence the decision of smallholder farmers to commercialize (or not) vegetables in the market. The probit regression model was run to find out why some smallholder farmers participate in the market and others do not. Table (7) shows the result of the probit regression model on the factors influencing smallholder commercialization of vegetables. The existence of a multicollinearity problem was checked for continuous variables. The values of Variance Inflation Factor (VIF) for each of the continuous variables were found to be less than 10; hence, there is no multicollinearity problem among all the hypothesized variables. The decision to participate in the vegetable products market was estimated by the maximum likelihood method. The marginal effect was used to explain the result as coefficients of the probit model are difficult to interpret since they measure the change in the unobservable y* associated with a change in one of the explanatory variables (that is, not partial effects). The chi-square tests applying appropriate degrees of freedom indicate that the overall goodness-of-fit of the probit model are statistically significant at 1% probability level.

3.2.1. Factors affecting market participation among smallholder vegetable producers

Out of the eleven explanatory variables hypothesized to have an influence on smallholder vegetable commercialization; six variables (age of household head, household family size, educational level, labour availability, market information and distance from market place) were found to be statistically significant in the study area (Table 7). On the other hand, demand for vegetables, transport access, off-farm income, extension and storage facility did not have a significant effect on farmers’ market participation. The data were analyzed and post estimation of the selection equation results was done to obtain the marginal effects. The marginal effects were used for interpretation since the coefficients of the selection equation have no direct interpretation. The reason is that they are just values that maximize the likelihood function. Marginal effects have a direct interpretation (Heckman, 1979). The discussion and interpretation of the significant explanatory variables in the probit model estimation are presented in (Table 7).

Table 7. Results of probit model for market participation of smallholder vegetable producers

Variables	Maximum likelihood (Coefficients)	Marginal effects (dy/dx)	P > z
Age of household heads	−.03408***	−.01304	0.001
Famly size	−.30972***	−.11857	0.000
Educational level	.41576*	.15917	0.091
Labour availability	.58227*	.22292	0.052
Access to market information	.74150**	.28388	0.015
Demand of vegetables	.17127	.06557	0.578
Access to transport services	.28282	.10827	0.329
Participation in off-farm activities	.14747	.05645	0.633
Access to extension service	−.41886	−.16036	0.172
Distance to market place	.04937*	.01890	0.083
Perishability of vegetable	.20919	.08008	0.486
_cons	1.6654	−.01304	0.016

Source: Model result (2017), ***, ** and * implies statistically significance at 1, 5, and 10% probability level respectively, N = 240, LR chi2 (12) = 223.67, Prob > chi2 = 0.0000, Log likelihood = −54.49, Pseudo R2 = 0.67.

3.2.2. Age of household heads

Age of household heads negatively influenced farmers’ likelihood to participate in vegetable marketing and significant at 1% probability level. An increase in the age of household head by 1 year decreases the probability of participating in vegetable marketing by 1.3%, all other factors held constant. This implies that younger people are more attached to technology and update their business minds with marketing issues so that youths were more participating in vegetable marketing than elders in the study area. This goes with the findings of Geoffrey et al. (2014) which stated that the younger people are more enthusiastic to participate in the pineapple market than the older people are. Barrett (2007) also stated that younger people participated more in the market because they are more receptive to new ideas and are less risk-averse than older people.

3.2.3. Family size

Household family size was found to have a negative and significant influence on farmers’ likelihood to participate in vegetable marketing at 1% probability level. The marginal effect shows that an increase in the size of the family decreases the probability of participation in vegetable marketing by 11.9%, all other factors held constant. This implies that the size of the family does not go along with the vegetable consumption level of the family because those households who have a number of children and do not have an excess of vegetable beyond of the consumption level they go for vegetable consumption rather marketing so that their participation in market output becomes very low. The finding also concurs with that of (Alene et al., 2008) which stated that a larger household is likely to consume more output, leaving smaller and decreasing proportions for sale.

3.2.4. Educational level

The educational level of farmers had a positive and significant (P ≤ 0.01) effect on smallholder farmer’s decision to enter the vegetable market. The marginal effect shows that an increase in the educational level of the farmers increases the probability of participating in smallholder participation of vegetable marketing in the output market by 15.9%, all other factors held constant. This implies that as the educational level of the farmers increases, their ability to get information on how to produce and sell vegetable increases (i.e., more farmers educated, they were more likely to produce and sell more at market price).

3.2.5. Labour availability

Labour availability was found to be positive and significant at the 10% probability level effect on smallholder farmer’s decision to enter the vegetable market. The marginal effect shows that an increase in the labour market of smallholders increases the probability of participating in smallholder participation of vegetable marketing in the output market by 22.3%, all other factors held constant. This indicates that large labour availability supplies more labour that can be well utilized in distributing vegetables to marketing places that offer a relatively better price.

3.2.6. Access to market information

Access to market information was found to influence smallholder farmers’ decision to sell their vegetables positively and significant at the 10% probability level. The marginal effect shows that an increase in getting information on time increases the probability of farmers’ participation in the marketing of vegetables in the output market by 28.4%, all other factors held constant. This indicates that farmers need to be able to get their products to market and receive equitable price treatment. Farmers need information pertaining to output prices so as to make the right decision, ahead of the production season, regarding which type of vegetable to produce and sell at the market. This result similar to the argument Goetz (1992) stated that better access to market information significantly raised the probability of market participation of households.

3.2.7. Distance to market place

Distance of market place was found to be a negatively and statistically significant influence on the commercialization of vegetables at a 1% probability level. The marginal effect estimates indicate that an increasing distance of farmers’ residence from the market place decreases the probability of participation in vegetable marketing by 1.9%, all other factors held constant.

4. Conclusion and policy implications

In Ethiopia, indigenous vegetables play a great role in the household economy. Southern parts of the country particularly Jimma zone have a good potential in vegetable production but with less market-oriented activities. Despite the production potential and importance of vegetables, there has been limited study on commercialization of vegetables. Vegetables are commodities that have a higher value in the domestic market. However, vegetables are more often used for consumption purposes than commercializing, and this is due to a lack of market information and other related factors. Although Ethiopian farmers are producing more of surplus vegetables, they are not much linked with markets and thus why their opportunity to diversify their livelihoods from vegetable production is very much limited. The finding indicated that smallholder vegetable farmers’ market participation is positively influenced by access to market information, education level of the head, labour availability, and proximity to market centers. On the contrary, the age of household head and family size reduces the likelihood of market participation. Therefore, based on the findings, the following recommendations were made in the commercialization of vegetables:

Expanding village markets in the rural areas and creating a good environment for older farmers as they can sell their vegetables at farm gate and village markets throughout brokers’ influence with a fair price. For those producers who had the problem of lack of information, equipping them with training on how to sell, where to sell and when to sell their products might be provided. The government should increase the marketing information and abilities of smallholder vegetable farmers especially on disseminating price information through rural community radios. Farmers with more children tend to less participate in vegetable marketing. Thus, access to appropriate family planning methods may help them to balance family size and participate in marketing.

Maintenance of transport, storage and other handling facilities is generally poor in the study area. Rising awareness on installing adequate storage facilities and proper regulation of temperature, humidity, air circulation, proper stacking pattern, regular inspection, and prompt produce disposal is a necessary step to improve farmer's market participation. Our study is limited to household-level factors affecting vegetable commercialization. Hence, there is a need for further research to critically analyze other factors affecting the commercialization of smallholder vegetable producers at community and macro-levels.

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Guta Regasa Megerssa

Guta Regasa Megerssa is a lecturer and researcher in the college of agriculture and veterinary medicine at Jimma University, Ethiopia. He received his MA in Development Studies from Mysore University in 2011. He offered development-related courses and conducted a few problems solving researchers for the community. His area of research interest is; commercialization, food security, agribusiness, diversification, enterprise, gender-related issues, sustainable development, rural development and innovation and so on.

Guta Regasa Megerssa