INTRODUCTION

In Ghana, tomato is the major vegetable consumed and used in almost all food preparations as one of the main ingredients. The crop alone constitutes about 38% of the vegetable budget in an average Ghanaian home (Agyekum, 2015). Currently, Ghana is the leading consumer of fresh tomatoes in Africa and the second in the world (Tuffour, 2018). Commercial production is mainly in the Upper East, Brong Ahafo, Northern and Greater Accra regions of Ghana. Production is also carried out mainly on family land or rented land of relatively small sizes (1-3 acres) (Monney et al., 2009). Varieties commonly grown in Ghana include Roma VFN, Pectomec VF, Tropimec, Rion Grande, Jaguar, Lindo, Titao Derma, Ada Cocoa, Laurano, Raki, Choco TP, Power, Reno, Rasta and Italy Heinz (MoFA, 2011). Production of tomato in Ghana is highly seasonal, reflecting differences in access to water and rainfall patterns (Robinson & Kolavali, 2010). Furthermore, tomato production is a major source of employment and income to both rural and urban dwellers as it is grown in many rural areas as well as the outskirts of town and cities to be supplied fresh to the urban markets and for exports (Ghanaveg reports, 2014).

In Ghana the total consumption of tomato outweighs its production as a result, the country is a net importer of tomato even from neighbouring countries like Burkina Faso. Tomato production in Ghana has not been able to reach potential yields compared to other countries, and hence has not been able to produce to sustain existing processing plants, as well as improving the livelihoods of farm households and other stakeholders in the tomato value chain. Ghana continues to spend scarce foreign exchange to import fresh tomato from Burkina Faso and processed tomato produced from China and other parts of the world. This import bill keeps increasing over the years with increase in population despite the country’s potential and comparative advantage to produce the crop locally. In spite of the benefits associated with improved seeds in terms of yield and quality, smallholder farmers in Ghana still cultivate local varieties which are often of low quality (Robinson & Kolavalli, 2010). This is because, for the past decade there has not been any improved tomato variety released. Most of the varieties cultivated are local varieties which have short shelf life with high perishability and low dry matter contents. This makes tomato production less profitable and hence farmers do not engage in its production. Subsequently, domestic demand outweighs supply resulting in high import of tomato to meet the increasing local consumption. The need for improved tomato varieties is vital given the increasing dependent of the country on large quantity of imports resulting in high import bills on Ghana’s economy.

In view of this, the Korea Programme on International Agriculture (KOPIA), Ghana Center together with Council for Scientific and Industrial Research - Crops Research Institute (CSIR-CRI) aimed at increasing tomato production in Ghana through the provision of high yielding varieties, better agronomic practices and improved technologies to help boost tomato yields, cut down on Ghana’s import bill on tomato and ultimately, improve wellbeing of farmers. The release and subsequent adoption of improved tomato varieties with good adaptability to local conditions coupled with good cultural practices has the potential of increasing tomato productivity, increasing incomes of farmers and reducing the high import bill for tomato in Ghana.

This paper evaluates farmers’ perceptions and preferences for improved varietal traits among smallholder tomato farmers in Ghana. It further assesses the constraints in tomato production and analyses farmer preferences for improved tomato varietal traits, and estimates the factors influencing the number of improved traits in an improved tomato variety. Such information is vital for tomato research and development in Ghana. The results from this paper presents vital insights for research for developing demand-driven improved tomato varieties to enhance adoption. Furthermore, it will contribute to the existing literature on tomato production by presenting insights on the relative importance of the major attributes that accrues to an improved tomato variety suitable for smallholder tomato farmers in Ghana. Ultimately, this will result in increasing the productivity of the entire industry to meet local demand, reduce the escalating import bill and reduce poverty among tomato producing households in Ghana.

METHODOLOGY

Study area

The study was carried out in the Offinso Municipality of Ashanti Region and Wenchi Municipality in the Brong Ahafo region of Ghana. These municipalities are the hub for tomato production in Ghana. The majority of the tomato produced in Ghana can be traced from these municipalities.

The Offinso North Municipal is located in the northwestern part of the region and shares common boundaries with Techiman, Sunyani, Tano and Nkoranza districts of the Brong Ahafo region. Agriculture is the major economic activity of the people. The major crops cultivated are maize, plantain, cassava, yam and vegetables especially tomato. Its district capital is Akomadan and other major settlements include Nkwakwaa, Asempanaye, Asuoso, Nkenkaasu and Afrancho.

The Wench Municipal is located in the western part of the Brong Ahafo region and bounded by the south by Sunyani Municipality and to the North by Kintampo South District. It is also shares common boundaries with Tain District to the West and Techiman Municipal to the East. Agriculture is the major economic activity of the people in these Municipalities with tomato production being the major agricultural economic activity. Major tomato growing communities include Akrobi, Amponsakrom, Asuogya, Awisa, Badu, Nchiraa, Nsawkaw, Tromeso and Subinso. These communities were randomly drawn from a list of major tomato producing communities in each Municipality.

Data and sampling

The data used in this paper was obtained through multistage sampling technique collected from a total, three hundred and six (306) smallholder tomato farmers who were randomly selected from ten (10) randomly selected major tomato growing communities in the Wenchi Municipal and five (5) randomly selected major tomato growing communities in the Offinso North Municipal. Primary data was collected using structured questionnaires through individual interviews.

Analytical Framework

The data was analysed using Stata software version 15 and SPSS software version 25. Descriptive statistics such as mean, standard deviation and percentages presented in graphs and tables were used to summarize the data. Additionally, the chi-square test of association and the student t-test were used to test the relationships between various socio-demographic, institutional and farm-level factors. Furthermore, a 5–point Likert scale using the following perception indices; 5 = strongly agree; 4 = agree; 3 = neutral: 2 = disagree; 1 = strongly disagree, was used to examine farmers’ perceptions on improved tomato varieties. The mean perception index which is the mean of each of these indices is used to evaluate the perceptions of the farmers regarding each of these attributes.

Ranking of constraints and varietal attributes

Following Mattson (1986), the Kendall’s coefficient of concordance (W) was used to rank the constraints to tomato production as well as improved varietal attributes preferred by farmers. The Kendall’s W was computed as:

Where:

Subsequently, a lower mean rank indicates that the constraint is more important. The Kendall’s W indicates the level of agreement among the farmers of the rankings obtained. Appropriately, a higher Kendall’s W denotes high level of agreement on the rankings.

Estimating the relative importance of the attributes in contributing to the preference and rankings

To evaluate the relative importance of the attributes, we estimate a Relative Importance Index (RII). The RII is the mean for an attribute which gives it weight in the perceptions of respondents. Subsequently, the factor with the highest weight obtains an RII = 1, while the next factor with lower weight obtains an RII = 2, and so on. The weighting is thus computed by the summation of the Likert allocation divided by number of responses. Thus, the RII for attributes for improved tomato varieties is expressed as:

where w is the weighting given to each attribute by the respondent and ranges from 1 to 4. for this analyses, n₁ = number of respondents who indicated that an attribute is of No Importance, n₂ = number of respondents who indicated that an attribute is of Neutral Importance, n₃ = number of respondents who indicated that an attribute is Important, while n₄ = number of respondents who indicated that an attribute is Very Important. A is the highest weight which corresponds to 4 in this paper, and N is the total number of respondents. Accordingly, the estimated relative importance index ranges from 0 to 1 (Tam and Le, 2006).

Examining the factors influencing the number of attributes preferred by farmers in improved tomato varieties

In examining the factors affecting the intensity of farmers’ preference for an improved tomato variety the Poisson regression model was employed. This model has extensively been applied to explain and predict the probabilities of the number of times an event occurs or is chosen (Ramirez and Shultz, 2000; Rahelizatovo and Gillespie 2004; Kim et al., (2005). In this paper, the Poisson regression model is used to analyse the factors influencing the number (count) of attributes a farmer expects in an improved tomato variety. The dependent variable of the model (y) is a count of the number of attributes or characteristics a tomato farmer prefers to see in an improved variety; that is,

Y is taken as a Poisson random variable; hence its probability density function can be represented as:

where У₁ is the number of attributes chosen by a farmer and x_i are determinants of improved variety preference. The probability density function (1) has one parameter λ which is the mean (and variance) of Y. That is; E(Y) = var(Y) = λ.

Therefore, the parameter λ_i usually take a log linear functional form defined as:

Hence the log likelihood function for the Poisson regression model is therefore given as:

RESULT AND DISCUSSION

Characteristics of tomato farmers in the Wenchi and Offinso North Municipalities

Table 1 presents the summary of socio-demographic characterises of sampled tomato farmers in the study location. The result shows that the mean age of the farmers is 44 years and is consistent across the districts. This implies that, the population of tomato farmers in the selected districts are of a youthful age and hence it has a positive effect on the prospects of tomato production in Ghana. It further implies that, the youth are venturing into tomato production, targeting development interventions at the youth have the tendency of producing the needed results. The average years of schooling of a typical tomato farmer in Ghana is 7. However, the Wenchi recorded a higher average of 9 years whiles that for the Offinso North is 6 years and is significant at the 1% level. This suggests that the majority of the farmers are educated up to the basic level of formal education. On average tomato farmers in the study location have attained about 20 years of experience in farming and 15 years of experience in tomato farming. These values are consistently greater in Offinso North than in Wenchi.

The result further shows that, majority of tomato farmers depend on their tomato farms as the main source of livelihood as the minority of them are engaged in off-farm activities with mean off-farm income per year which is a little above GHC 3000 and was significantly greater in Wenchi than in the Offinso North. A typical tomato producing household in Ghana has high household size of 8 members depicting a high household labour availability. Tomato farming in Ghana constitutes largely of men (Table 1).

Figure 1, shows the distribution of educational level of farmers by Municipality. The figure shows that the illiteracy rate among tomato farmers is generally high. The majority (68%) of the farmers were educated only up to the basic education with about 17% who had no formal education. Only 2% of the farmers were educated up to the tertiary level. Educated farmers are better able to understand and adopt technologies (Muange et al, 2014). In this regard there is the need for policies that will encourage formal education among famers.

Table 2 presents the distribution of household head by Municipality. Generally, household heads are the final decision makers in most farm households in Ghana. Similarly, the sex of the household head influences adoption decisions in the household (Kabunga et al., 2012; Donkoh et al., 2014). The result shows that, largely, household heads are dominated by men with only 16% of households being headed by women. With respect to the Municipalities, 71.7% of households in the Offinso North and 94% households in the Wenchi were headed by men. Women-headed households were observed to be 20% in the Wenchi Municipal and 11% in Offinso North. The above result indicates that men are dominant household heads in the study location.

Farm-level characteristics

Table 3 presents the summary of farm-level characteristics of sampled farmers by Municipality. On the average, a typical tomato farming household in Ghana consist of three (3) members. This value is relatively low compared to the average household size of eight (8) members. A tomato farmer commonly owns a single plot of land of about two (2) acres and cultivates once in a year, essentially due to the lack of irrigation systems to permit two (2) growing seasons in a single year. This means the majority of the farmers depend solely on the rains for cultivation hence in the presence of irregular rains or lack of rains an improved variety which is drought resistant and adapted to the local climate will be essential. This is confirmed in Figure 2 which shows the type of production systems adopted by farmers. In general, the majority (67%) of the farmers depended on the rainfall for their tomato cultivation.

Training on improved tomato production techniques

Training of farmers is an essential precursor to technology adoption as it enables farmers to adopt and to realize the full potential of the improved or new technology. Figure 3 presents the training received and training needs of tomato farmers. The result shows that, the major training need among tomato farmers is in the area of pest and disease management followed by variety/seed selection, and requisite skills in post-harvest management.

However, most of the farmers have already received training in good agricultural practices. The results suggest that any form of capacity building for tomato farmers to enhance productivity of tomato in Ghana and particularly in the Wenchi and Offinso south municipality should focus on pest and disease management, varietal selection and post-harvest handling.

Farmers preferences for varietal attributes in an improved tomato variety

Table 4 presents farmers’ preference for improved tomato varieties. The result shows that almost all the framers prefer to have improved/new tomato varieties (98%). This indicated a strong demand for improved tomato varieties in Ghana and hence presents an opportunity for research to identify the relevant traits in developing such improved tomato varieties that will meet the needs of farmers.

Ranking of farmers’ preferred attributes of an improved varietal attributes

Table 5 shows the ranking of attributes preferred by the farmers in an improved tomato variety across the two municipalities. The estimate for the Kendall’s W for the pooled (overall) is 0.607 which implies that there is 60.7 percent agreement among the farmers of the rankings of the preferences presented in the table strongly justifying the rankings of the preferred attributes presented by the farmers. Still with regards to the pooled sample, the three (3) most important attributes preferred are varieties with long shelf life, large fruit size and improved fruit quality.

Tomato is a highly perishable crop and noted for its short shelf life so farmers find it necessary to have fruits that can be stored for extended periods of time without deteriorating as this tends to reduce wastage thereby reducing postharvest losses. Large fruit size also obtained a mean rank of 2.70 whereas improved fruit quality scored a mean rank of 3.62 making them the second and third most preferred traits, respectively. This indicate that farmers preferred attributes that directly affect the fruit of the tomato. That is, having a longer shelf life, improvement in fruit quality and an increase in size of the fruit. The least preferred improved varietal attribute again with regards to the overall sample was tolerance to water logged areas with a mean rank of 8.79 and represented the least (10^th) rank. Implying that, an improved variety which is tolerant to a water logged area is least important attribute farmers look out for in an improved variety hence not an essential breeding objective for improved tomato variety development. This can be ascribed to the fact that study locations do not have serious flowing issues. Accordingly, breeding objectives needs to focus on breeding for increasing shelf life, size and fruit quality to enhance the uptake of improve tomato variety in Ghana.

However, there are differences in the rankings of varietal attributes across two municipalities. Whereas farmers in Wenchi ranked improved fruit quality as the most desired attribute this trait was ranked as the fifth (5^th) desired attribute by farmers in Offinso who rather ranked long shelf life as the most important varietal attribute. Again, whereas non-staked was ranked least preferred trait at Offinso, it was ranked 7^th important trait by farmers in Wenchi who ranked tolerance to water logged areas as the last (10^th) trait. Furthermore, the variety’s resistance to pest and disease was very important to farmers in Offinso as they ranked it third (3rd), however famers in Wenchi considered it as the 5^th important varietal trait.

Relative importance analyses of the attributes farmers prefer in an improved tomato variety

Table 6 presents the results from the relative importance analyses of farmer preferred attributes in an improved tomato variety. The attributes can be grouped into five (4) main categories namely marketability (fruit size and quality), storability (shelf life), resistance to natural/biological hazards (tolerant to pest and disease and drought, nonstaked and early maturity) and adaptation to microclimate conditions (poor soils, extreme heat and water logging). The results shows that, long shelf life had the highest relative importance index (0.912). The implication is that farmers consider long shelf life as the most important attribute in an improved tomato variety. The relative importance index is close to 1 which indicates the importance farmers place on shelf life in tomato varietal selection. Close to 67% of the farmers responded to long shelf life as very important varietal attribute and this explain why the attribute had the highest relative importance index.

As shown by the rankings in Table 5, the place and time utilities of tomato are affected directly by its perishability. In the absence of adequate storage facilities or cold storage, farmers tend to prefer tomato varieties that can last much longer after harvest in order to minimize post-harvest loses and maximize revenues from the sales of tomato. The importance of shelf life in tomato varietal section and evaluation is consistent with findings of Tewodros & Negasi (2014). The next most important attribute is fruit quality. This attribute had relative importance index of 0.904. About 65% of the farmers responded to fruit quality as very important attribute and this comes after shelf life in order of preference. Fruit quality is one of the most important attributes that consumers look at for in the market places. Good looking and fresh tomatoes usually attract price premiums.

That is the market potentials of freshly harvested tomato is affected by fruit quality and profit maximizing farmers tend to place much premium on fruit quality when it comes to varietal selection. Size of the tomato fruit follows its quality in terms of relative importance. The relative importance index for fruit size is 0.879. About 50% of the farmers indicated that quality of fruit was very important in varietal selection and evaluation. Consumers usually prefer good looking and sizeable tomato fruits to bruised and unattractive once. It must be emphasized that freshly harvested tomatoes are sold in carts/boxes. Number of tomatoes fruits per cart is reduced for sizeable varieties as compared with smaller once and farmers will go for varieties that offer minimum fruits per cart. Tolerant to pest and diseases, tolerant to drought and early maturity had relative importance indices of 0.867, 0.815 and 0.872 respectively and these fall within the very important category. Natural and biological hazards such as flooding, drought, strong wind, pest and disease tend to impact negatively on the performance and contribute significantly to yield loses in tomato production. Farmers therefore consider varieties that can resist extreme climate change events. Tolerant to extreme heat, soil fertility and water logging had relative important indices of 0.788 and 0.716 respectively. Most arable lands in Ghana are characterized by poor physical and chemical properties including low fertility and aeration and this is largely attributed to poor soil management practices. Smallholder farmers prefer tomato varieties that can adapt to the prevailing soil conditions. Staking is a labour intensive activity in tomato production and it increases the production cost for farmers so farmers will prefer non-staked tomato varieties to avoid or minimize production cost. Bertin and Brunel (2003) and Tigist (2008) have highlighted the importance of marketability, growth characteristics and resistance to natural hazards and adaptation to unfavourable soil conditions in tomato varietal selection.

Factors influencing the number of varietal traits desired in an improved tomato variety

The demand or adoption of improved crop varieties depend largely on its attributes, Rogers, (2003). Farmers are likely to adopt new crop varieties whose attributes they consider favourable to their prevailing farming circumstances. Its therefore not surprising that factors that influence the adoption of improved crop varieties tend to favour the acceptability of their attributes. Poisson regression model was used to estimate the factors influencing the number of varietal traits a farmer expects in an improved tomato variety. The results are presented in Table 7. Experience in tomato cultivation had a positive significant effect on the number of attributes preferred in an improved tomato variety. Farmers with more year of experience in tomato cultivation tend to appreciate the essence of certain traits in an improved variety and hence are more likely to prefer more attributes. Sex of farmer was statistically different at the 5% level implying that the number of attributes preferred in an improved tomato variety is significantly more for men than women farmers. Further, the results show that, educated farmers preferred more attributes in an improved tomato variety. This is consistent with Kabunga et al., (2012) who posited that the level of the education also contributes significantly in the adoption process because the higher the educational level easier to adapt to the use of the improved varieties of seeds.

Farmers with lager household sizes chose and expect more attributes in an improved variety. The results further show a positive effect of household size on the number of attributes preferred in an improved tomato variety. Thus the larger the size of the farm household more likely they are to prefer more attributes in an improved tomato variety. This results resonates with Asante et al. (2014) who found technology adoption to be influenced by large household size due to the availability of family labour though the large household members. Farmers who belonged to an FBO and had extension contacts preferred more attributes than to those who do not at the 10% and 5% significance levels, respectively. Both the FBO and extension provides opportunity to farmers to be educated through trainings and farmer education which enhances their knowledge hence not surprising such farmers prefers more attributes in an improved tomato variety. Further, farmers with larger farm sizes and off-farm income prefer more attributes probably due to the availability of the relatively greater resources at their disposal in terms of land and capital.

Constraints in the tomato production industry

Table 8 depicts the ranking of the constraints faced by farmers in tomato cultivation. Access to capital is the major challenge faced by tomato farmers. This was followed by low price of the commodity and the limited demand for the fruits. High postharvest loses and difficulty in accessing improved tomato varieties were ranked 4^th and 5^th , respectively. The least constraint observed was limited access to tractors for land preparation probably due to the fact that majority of the famers cultivated the crop on small to medium scale on an average of a two (2) acre land and hence mechanical land preparation not being a major priority. Similar ranking as in the pooled sample was observed among famers in the two (2) districts with no differences. However, per the Kendall’s W values, the agreement in order of raking was much higher in Offinso than in Wenchi with vales of 0.782 and 0.587 respectively. The value of the Kendall’s W for the total sample was 0.653, indicating a 65.3% agreement among the farmers with respect to the constraint rankings order and is statistically significant.

Intensity of constraints in tomato production

The results in Table 8 focused on the rankings of the constraints farmers are facing in tomato production, however, this did not reveal the intensity of each constraint in terms of how severe they affect the farmers. Accordingly, the extent of severity of these constraints is presented in Figure 4. The Figure shows that, in terms of lack of access to capital, 85 percent of the farmers perceived that the lack of capital is a very severe constraint with only 1.96 percent who attested that, to them capital is not a constraint. Similarly, the majority, 80% and 74% attested respectively that low pricing and low demand of their produce is a very severe challenge they face. Further 57% of the farmers who attested to postharvest loses being a challenged said it was very severe leaving the remaining 36% and 8% who said it was severe and was not severe respectively. Lack of technical know-how and access to improved varieties were perceived as having very severe effects on their production (42.48% and 43.46%, respectively). It is again confirmed that the least ranked constraint; limited access to tractors for land preparation is seen as not having a serious effect on the farmers as indicated by the 2nd majority of the farmers constituting 36 percent.

Conclusions and Policy Recommendations

This paper examined farmers’ perceptions and preferences for improved varietal traits in the Wenchi and Offinso North Municipalities of Ghana. Data was analyzed using perception indices, Kendall’s Coefficient of concordance and the Poisson regression models. The results shows that tomato farmers are willing to adopt an improved variety that possess the attributes including long shelf life, good fruit quality and large fruit size. A typical tomato farmer cultivates an average land size of 2 acres and production is mainly rainfed, as a result, farmers cultivate the crop once in a year as majority of the farmers do not have access to irrigation. The varietal attribute that is most desired by farmers is shelf life followed by fruit quality and fruit size. The study further revealed that tomato farmers do not have adequate skills in pest and disease management as the majority of them indicated that as a key training need. Training programs organized for tomato farmers should thus focus more on pest and disease management. The number of varietal attributes preferred by farmers is positively influenced by sex of farmer, education, household size, access to credit, FBO membership, extension contacts farm size and off-farm income. These factors and attributes need to be carefully considered by breeders and policy makers in the development of an improved tomato variety. The major constraint identified in tomato production is limited access to capital, low commodity prices and low demand for the fruits. The potential of adoption of improved tomato varieties in Ghana will be augmented through paragenetic measures towards addressing these constraints.

적 요

이 논문은 가나의 웬치 및 북 오핀소 군의 개량품종 특성에 대한 농민들의 인식과 선호도를 조사 분석한 것이다. 무작위 로 선택된 306 명의 토마토 생산 농가의 데이터를 인식지수, Kendall의 일치도 계수 및 포아송 회귀 모형을 사용하여 분석 했다. 분석결과 토마토 생산 농민들은 유통 기한이 길고 과일 품질이 우수하며 과일 크기가 큰 개량 품종을 채택하고 있음 을 보여준다. 또한 이 연구는 토마토 농민이 해충 및 질병 관 리에 필요한 기술이 부족함을 보여주며, 농민들이 선호하는 품 종 특성은 농민의 성별, 교육, 토마토 재배 경험, 가구 규모, 신용 기회, 농민조직 가입, 시범농장 규모 및 농외 소득에 긍 정적인 영향을 받았다. 육종가와 정책 입안자들은 개선된 토 마토 품종을 개발할 때 이러한 요소와 속성을 신중하게 고려 해야 할 것이다. 토마토 생산에서 확인 된 주요 제약은 자본 에의 접근 제약, 낮은 토마토 판매가격, 경지 준비를 위한 트 랙터 이용 애로 등이 포함된다. 가나에서 개량 된 토마토 품 종의 채택 가능성은 이러한 제약을 해결하기 위한 조치를 통 해 증대 될 것이다.