Introduction
In recent years, farm prices and income levels, financial problems and uncertainty of government programs have encouraged farmers to seek alternative marketing strategies (Cottingham et al., Reference Cottingham, Hofland, Lenon, Roper and Techtmann1994). Pressures on farming in the form of collapsing prices, overproduction, market saturation, internal and global competition, concentration of powerful buyers and agricultural policy reform are pushing interest in alternative marketing into the mainstream (IATP, 1998). One popular marketing option that allows farmers to receive a higher return for their crops is direct marketing (Young, Reference Young1995).
Direct marketing includes any marketing method whereby farmers sell their products directly to consumers (Bruch and Ernst, Reference Bruch and Ernst2010). Direct marketing eliminates the ‘middle man’ functions of food exchange, transportation, storage and processing (Cottingham et al., Reference Cottingham, Hofland, Lenon, Roper and Techtmann1994). Several reasons account for the increased interest in farm direct marketing. One is dissatisfaction with farm commodity prices. The farm price is often only a fraction of the retail food price. Prices for produce sold directly to consumers can be substantially higher than typical wholesale prices. Another reason is that producers value the relationships they form with the consumers as well as the opportunity to receive immediate feedback on their products. Consumers value the fresh, quality products along with the opportunity to support local producers (Alberta Ag-Info Centre, 2013). The direct sale of agricultural products plays an important role in the diversification of the activities of the rural population and helps to develop rural areas. Not only does this form ensure the provision of healthy food for consumers, but it also provides increased income to farmers, allowing farmers to remain in the countryside (Fehér, Reference Fehér and Adisa2012).
Many direct marketing options are available. Selecting one depends on the individual, the farm location, the volume of products to sell and other factors. Major direct marketing options include roadside stands, U-pick operations, roadside markets and farmers’ markets. Other options include ‘honor selling,’ gift baskets and mail-order or Internet sales, community-supported agriculture, selling direct to restaurants and other food institutions, agritourism, ‘peddling,’ and rent-a-tree or rent-a-plot arrangements (Burt and Wolfley, Reference Burt and Wolfley2009).
Some statistical data reveal that the importance of direct marketing for farm products has increased in recent years. For example, throughout the USA, 144,530 farms sold US$1.3 billion in fresh, edible agricultural products directly to consumers in 2012. This was a 6% increase in farms and an 8% increase in sales over 2007, the last time the agriculture census was conducted. Farms with direct sales to consumers made up 6.9% of the nation's 2.1 million farms in 2012 (USDA, 2014). According to a study carried out by the UK's University of Essex for the Soil Association, organic farms in the UK and Ireland involved in ‘on-farm processing’ and direct marketing enterprises employed 64% more people than organic farms without such activities, with a significant proportion (39%) of organic farms engaged in this type of business innovation (Hird et al., Reference Hird, Holder, Chandrasekaran, Benning, Schimpf and Becheva2010). In Austria, family farmers in particular are engaged in direct marketing and have built direct contacts with other stakeholders to get information and networks (Fujisawa et al., Reference Fujisawa, Kobayashi, Johnston and New2015). In 2007, for less than 0.2% of farms in France, Germany and the UK did direct sales account for more than half of the value of their output. Yet at the same time, there were 384,995 Polish farms (16% of the total) for which direct sales accounted for over half of the total sales value. Direct marketing of farm products is most developed in Italy, where the number of farms involved in such activities rose by 64% between 2001 and 2009 (Gorton et al., Reference Gorton, Salvioni and Hubbard2014).
This paper attempts to answer the basic question: What are the factors that drive farmers’ decisions to sell their produce in direct marketing? To answer this question, a range of variables, including demographic, socio-economic, market, and farm specific features, were jointly estimated in order to be able to evaluate the influence of each of them. As described in the literature review section, even though there are some studies done to explore factors that may be related to a farmer's decision to participate in direct marketing, further research is needed concerning regional differences affecting farmer decisions to participate in these activities. So far, no similar analysis has been carried out for Turkey. This paper aims to fill this gap in the existing literature. The study was based on primary data obtained in a cross-section survey of 102 cherry growers in order to determine the factors that affect farmers’ decisions to participate in direct marketing in Turkey. In addition, this study also explores what options of direct farm marketing are being used by farmers and their future plans for direct marketing strategies (DMSs). Lastly, in this study, further suggestions are put forward on how to improve the use of DMSs by farmers in Turkey.
Literature Review
There is considerable literature that analyzes characteristics and attributes of consumers who purchase products through direct-to-consumer (DTC) outlets. On the other hand, there are relatively few studies that focus on producer behavior and characteristics regarding DMSs (Monson et al., Reference Monson, Mainville and Kuminoff2008; Wolanin, Reference Wolanin2013; Park et al., Reference Park, Mishra and Wozniak2014). There is no study that has actually been carried out to determine the factors that influence farmers’ decisions to participate in direct marketing in Turkey. However, few studies that have been carried out to date have analyzed the factors that influence farmers’ decisions to participate in direct marketing (Gilg and Battershill, Reference Gilg and Battershill1999; Monson et al., Reference Monson, Mainville and Kuminoff2008; Aguglia et al., Reference Aguglia, Francesco and Salvioni2009; Auld et al., Reference Auld, Thilmany and Jones2009; Corsi et al., Reference Corsi, Borsotto, Borri and Strøm2009, Reference Corsi, Novelli and Pettenati2014; Nyaupane and Gillespie, Reference Nyaupane and Gillespie2010; Timmons and Wang, Reference Timmons and Wang2010; Detre et al., Reference Detre, Mark, Mishra and Adhikari2011; Park et al., Reference Park, Mishra and Wozniak2011, Reference Park, Mishra and Wozniak2014; Uematsu and Mishra, Reference Uematsu and Mishra2011; Sage and Goldberger, Reference Sage and Goldberger2012; Wolanin, Reference Wolanin2013; Benedek et al., Reference Benedek, Fertő, Baráth and Tóth2014; Park, Reference Park2015; Muthini, Reference Muthini2015).
Gilg and Battershill (Reference Gilg and Battershill1999) examined the household factors influencing the participation of French farms in direct selling of farm produce as an alternative farming strategy. While age, family life cycle, succession and labor relations were not found to be significant in understanding farm decision making, in this study farming background, education and attitudes toward profit were found to significantly influence decisions about direct marketing.
Monson et al. (Reference Monson, Mainville and Kuminoff2008) analyzed how farm size, the importance of high-value crops, organic production, experience and demographic factors affect a producer's reliance on direct markets. They found that farm size, high-value crop production, noncertified organic production methods and household size are determinants of the share of total farm output sold through direct marketing outlets. Results from this study suggest that producers who operate smaller farms, are less reliant on small fruit, implement organic production without USDA certification and live in smaller households are more likely to sell their products through direct outlets.
Aguglia et al. (Reference Aguglia, Francesco and Salvioni2009) focused on the determinants of the adoption of direct selling in Italian farms. They indicated that the probability of uptake of this marketing strategy is higher among multifunctional farms, for example, among organic farms. They then tested on a smaller regional sample the influence of proximity to urban areas on the adoption of direct selling. The results show that the probability of uptake is higher when the farm is close to big urban markets.
Corsi et al. (Reference Corsi, Borsotto, Borri and Strøm2009) analyzed the choices of marketing chains by organic farmers and identified the determinants of their choices. They found that farmers’ personal characteristics influenced their choice, and that more educated and skilled farmers were less likely to choose traditional marketing chains and more likely to engage in the new marketing chains. They pointed out that large farms choose traditional chains rather than the direct and the short chains. According to them, the other main determinant of the choice was the type of farming, with some types more suitable for the traditional chains, and other for the direct and short ones.
Auld et al. (Reference Auld, Thilmany and Jones2009) identified production characteristics that affect the relative propensity to sell locally and explore opportunities to support more direct and local marketing by producers based on these characteristics. The results of this study revealed that producer size, product perishability, proximity to a population center, consumer perceptions and producer philosophy influenced where Colorado growers sell. They concluded that small farmers sold the majority of their crops locally, the produce grown by small farmers in Colorado had significantly lower food miles than conventional sources. Finally, direct sales almost always result in local sales and, subsequently, lower food miles. The authors also pointed out that product characteristics also affected where produce was sold. Most growers believed that perishability was advantageous to market ‘local’ to maintain freshness, flavor and quality and reduce the need for additives.
Timmons and Wang (Reference Timmons and Wang2010) identified major factors associated with direct food sales across states and counties in the USA using 2007 USDA Census of Agriculture data. They used regression analysis to identify major factors that may explain variations in direct food sales. They found that four variables, average farm size, population density, region and available farmland, together explained most of the variation in direct food sales across states. Average farm size was negatively associated with direct sales, i.e., the presence of smaller farms was correlated to higher direct sales. Population density and farmland were reported to be positively correlated with direct food sales. The authors also revealed the wide regional differences in direct food sales.
Detre et al. (Reference Detre, Mark, Mishra and Adhikari2011) investigated the factors affecting the adoption of a DMS by farmers and its impact on the gross sales of farm operations in the USA. They found that the production of organic crops and the regional location of the farm positively affect adoption of a DMS. However, they highlighted that adoption of a DMS has a negative relation to large farms, farms with production contracts and farms specializing in cash grains.
Uematsu and Mishra (Reference Uematsu and Mishra2011) identified factors affecting the total number of DMSs adopted by farmers. Then they estimated a quantile regression model to assess the impact of the intensity of adoption of DMSs on gross cash farm income. They found that farming experience in years, total acres operated, direct payments received by the farm in dollars and farm location were all found to negatively affect adoption intensity of DMSs. They also found that farmers with livestock, high value and other field crops farms are likely to more intensively adopt DMSs than cotton and cash grain farmers. According to the authors, greater educational attainment, farming as primary occupation, farming as a primary occupation for the respondent's spouse, seeking advice from Natural Resource Conservation Service agents, receiving payments from the Conservation Reserve Program and having internet access at their farm were all positively associated with the adoption of DMSs.
Nyaupane and Gillespie (Reference Nyaupane and Gillespie2010) used the results of a 2008 survey data from the Louisiana crawfish industry to investigate the factors influencing farmer selection of a crawfish marketing outlet. Results from this study show that farm size, farm income, household income, age, education and pre-market grading and washing operations significantly affect farmer selection of marketing outlet. The authors pointed out that farmers who washed crawfish before selling and had higher percentages of their farm income coming from crawfish were the more likely farmers to market direct to consumers. Older, less highly educated farmers were more likely to market direct to processors. Scale of operation was the major determinant of whether farmers would sell directly to retailers, as larger farmers were the ones who had the volume required to sell directly to the retail market.
Sage and Goldberger (Reference Sage and Goldberger2012) investigated the effects of producers’ ‘worlds of justification’ on willingness to participate in DTC markets. They found that a producer who had always farmed organically was nearly three times more likely to participate in direct marketing compared with a producer who switched from conventional production. They revealed that dairy/livestock producers were significantly less likely to participate in a direct-to-consumer marketing strategy compared with other major producer types. The results also indicated that both the MI (Market Performance and Industrial Efficiency) and CG (Civic Equality and Green) factors were demonstrated in this model to be significant factors in the decision to direct market. Unit increases in the value placed on MI justifications for their decision to farm organically substantially reduces the likelihood of participating in direct market venues. Those producers who placed high value on CG justifications were significantly more likely to direct market than those who placed low value on CG justifications.
Park et al. (Reference Park, Mishra and Wozniak2011) investigated the factors affecting choices of direct sales by farmers. They determined that using the Internet for farm commerce and growing a diversified selection of products (more enterprises) increases the likelihood that a farmer uses DTC marketing outlets.
Wolanin (Reference Wolanin2013) focused on the factors affecting the intensity with which fruit and vegetable farmers incorporate DMSs into their agricultural enterprises, where intensity is measured by the percentage of sales made through DTC outlets. Wolanin found that the factors significantly affecting producer intensity of adoption of DMSs are age, use of university/extension publications, percentage of income from farming and access to food hub organizations. Results from this study indicate that farmers getting closer to their retirement age or slowing down in their off-farm jobs, using University/Extension publications to obtain information on how to better market produce in the past 5 yr, with a percentage of income from farming of 25% or less, or who have their farming operation located in a county with no access to a food hub organization tend to make a larger percentage of their fruit and vegetable sales through DTC outlets or tend to more intensively adopt DMSs.
Benedek et al. (Reference Benedek, Fertő, Baráth and Tóth2014) used an econometric model to identify factors that drive farmers’ decisions on where to sell their produce (whether to sell at traditional or farmers’ markets). Results from this study indicated that a relatively young, educated and innovative group of small-scale farmers were interested mostly in selling at the newly introduced farmers’ markets.
Corsi et al. (Reference Corsi, Novelli and Pettenati2014) investigated the determinants of the choice of farmers to sell their products directly to consumers. They concluded that operators’ and farm characteristics were found to affect the choice of selling directly, but rather weakly. The most important factors affecting these choices were farm location and, for on-farm direct sales, the complementarity with agro-tourism and recreational activities. Relative to plains, farms located in the mountains are 12.2% more likely to sell their products on the farm, and farms in hills 7%. If the farm has some agro-tourism, or recreational activity, the likelihood of selling directly on the farm is increased by 25 and 11%, respectively. This study revealed an interesting finding regarding the type of farming (TF). All specialized TFs had a lower probability to sell directly on the farm relative to the mixed TFs, taken as reference.
Park et al. (Reference Park, Mishra and Wozniak2014) investigated firstly the role of marketing and management skills in the choice of DMSs by farmers. Secondly, they assessed the impact of management and marketing skills on the farm financial performance. Findings from this study show that management and marketing skills significantly affect DTC sales. According to their findings, increasing management skills––more ways to control input costs––increases the likelihood that a farmer uses DTC marketing outlets. Additionally, a larger share of income coming from sales of vegetable, fruits and nursery enterprises is associated with a greater likelihood that a farmer uses DTC marketing outlets. They also find that beginning farmers are more likely to use DTC marketing outlets.
Park (Reference Park2015) examined the impact of participation in direct marketing on the entire distribution of farm sales using the unconditional quantile regression estimator. Park assessed how farmers’ sales were influenced by their involvement in direct marketing. Results from this study indicated that smaller operations were more severely impacted when farmers participated in direct marketing compared with larger operations. Results also indicated that experienced farmers involved in direct marketing were faced with larger sales declines compared with farmers who were not involved in direct marketing. Finally, Park pointed out that although participation in direct marketing was associated with a decline in farm sales, farmers who experienced growth in off-farm income were better able to withstand these sales declines.
Muthini (Reference Muthini2015) assessed the factors that influence mango farmers’ choice of market channels in Makueni County. Multinomial logit model was used to quantify the factors affecting channel choice. Results of the study showed that farmers sold to three major channels, which are brokers, exporters and direct market. Majority of the farmers (58%) sold to brokers, 30% to export, while the rest sold to direct market. The multinomial logit results showed that farmers who owned a vehicle, were closer to the tarmac road and had access to market information were more likely to sell to direct market relative to brokers.
Materials and Methods
Data
An examination of related research studies concerned with direct farm marketing in various countries shows that the main products sold through direct marketing channels are fresh fruits and vegetables. This has also been observed for Turkey. In this study, a survey of cherry growers was conducted in order to determine the factors that influence farmers’ decisions to participate in direct marketing in Turkey.
Data were collected in 2012 via personal interviews of a random sample of 102 cherry farmers in Izmir Province in Turkey. The survey was implemented in the district of Kemalpasa in Izmir Province, which is an important area for cherry production in Turkey (Adanacioglu, Reference Adanacioglu2013). Izmir cherry production was about 56,772 tons in 2012, according to the Izmir Directorate of Provincial Food, Agriculture and Livestock (MFAL, 2012). Kemalpasa provides 86.32% of Izmir's cherry production. In the choice of subdistricts where the survey would be carried out, the number of cherry producers and the amount of production were taken into account. In this way, the three subdistricts Bagyurdu, Oren and Yigitler were selected.
The sample size was determined by using the proportional sampling method (Newbold, Reference Newbold1995).
$$n = \displaystyle{{Np(1 - p)} \over {(N - 1)\sigma _{{{\hat p}_x}}^2 + p(1 - p)}}$$
where n = sample size, N = number of farms (2157 farmers), p = the percentage of farmers who grow cherries (taken as 0.50 to reach maximum sample size), and σ2 px = variance.
According to the proportional sampling method, with a 95% confidence interval and 9.5% error margin, the required sample size was found to be 102.
Empirical model and description of variables
An ordinal logistic regression model was used to identify the factors that influence farmers’ decisions to participate in direct marketing. Standard statistical texts describe several approaches to specifying (parameterizing) an ordinal logistic regression model. Here, we used the cumulative logit model, which is the most common form. A cumulative logit is defined for the probability of having an ordinal response less than or equal to k, relative to the probability of having a response greater than k:
$$\eqalign{{\rm logit}[P(y \le k) \vert x] &= \, \ln \left[ {\displaystyle{{P(y \le k) \vert x} \over {P(y \gt k) \vert x}}} \right] \cr & =\, \ln \left[ {\displaystyle{{P(y = 1 \vert x) + \cdots + P(y = k \vert x)} \over {P(y = k + 1 \vert x) + \cdots + P(y = K \vert x)}}} \right] \cr &=\, {\beta _{0(k)}} - ({\beta _1}{x_1} + {\beta _2}{x_2} + \cdots + {\beta _p}{x_p})} $$
For an ordinal variable with K categories, K − 1 cumulative logit functions are defined. Each cumulative logit function includes a unique intercept or ‘cut point,’ β 0(k), but all share a common set of regression parameters for the p predictors, β = (β 1,…, β p ). Consequently, a cumulative logit model for an ordinal response variable with K categories and j = 1,…, p predictors requires the estimation of (K − 1) + p parameters. The concept of a ‘cumulative logit’ is certainly more complex than that of a baseline category logit that is employed in the parameterization of simple logistic regression or even multinomial logit regression models. In actuality, though, it is simply an alternative way of parameterizing the model for estimating the probability that a response will fall in ordinal category y = 1,…, K (Heeringa et al., Reference Heeringa, West and Berglund2010).
The dependent variable, in this model, is the degree of farmers’ tendency toward direct marketing options in ordinal form. Farmer respondents were asked to indicate the extent of their preference (1 = not at all; 5 = a great deal) for each of the following direct marketing options: local vegetable and fruit markets/farmers’ markets, Internet or e-mail sales, roadside stands, on-farm sales and community-supported agriculture. Afterward, the final score was computed by averaging the responses to all of the items. The ordinal dependent variable used in the current study was categorized into three groups based on the final scores as follows: 1: (a weak tendency toward direct marketing) if the final score ≤2, 2: (a moderate tendency toward direct marketing) if the final score >2–3, 3: (a strong tendency toward direct marketing) if the final score >3. In the analysis, category 3 was taken as the event of reference.
Generally, previous empirical studies have used a wide range of variables to determine factors influencing farmers’ decision to participate in direct marketing. In accordance with previous studies, the following variables were included as explanatory variables in the model: age (AgeD1, D2, D3, D4), educational attainment of the farm operator (EducD1, D2, D3, D4), farm operator's experience in cherry growing (FexpD1, D2, D3), farm size (FarmSizeD0, D1), cultivated area with cherry (CoSizeD1, D2, D3), level of specialization in cherry production (GPVCherryD0, D1) and total annual farm income (IncomeD1, D2, D3, D4, D5). Unlike previous studies, two variables were also included in the model as explanatory variables, namely gross margin per hectare for cherry production (GrmD1, D2, D3) and number of marketing channels used by farmers (MchnD0, D1). Variable definitions, as well as descriptive statistics for the explanatory variables employed in the ordinal logistic regression model, are presented in Table 1.
Descriptive statistics for variables used in the empirical models.
Age is hypothesized to be positively correlated with the decision to participate in direct marketing. This hypothesis is consistent with the finding of Wolanin (Reference Wolanin2013) who concluded that farmers getting closer to their retirement age tend to make a larger percentage of their fruit and vegetable sales through DTC outlets or tend to more intensively adopt DMSs. Wolanin (Reference Wolanin2013) cites that as farmers get close to the retirement age they may have more time available to be devoted to farming activities, including marketing, and therefore may increase the adoption intensity of DMSs. This hypothesis was also supported by the finding of Nyaupane and Gillespie (Reference Nyaupane and Gillespie2010) who concluded that older farmers were more likely to market direct to processors.
Education is hypothesized to be positively correlated with the decision to participate in direct marketing. This is consistent with the hypothesis put forward by Monson et al. (Reference Monson, Mainville and Kuminoff2008) in their study. They put forward the hypothesis that as the primary decision maker's degree of education increases, reliance on direct marketing will increase. Wolanin (Reference Wolanin2013) cites that producers who have a bachelor or graduate degree are expected to have a larger percentage of their sales made through DTC outlets. This hypothesis was also supported by the findings of Gilg and Battershill (Reference Gilg and Battershill1999), Corsi et al. (Reference Corsi, Borsotto, Borri and Strøm2009) and Benedek et al. (Reference Benedek, Fertő, Baráth and Tóth2014).
Farm operator's experience in cherry growing is hypothesized to be positively correlated with the decision to participate in direct marketing. This is also consistent with the hypothesis put forward by Monson et al. (Reference Monson, Mainville and Kuminoff2008) in their study. They put forward the hypothesis that producers who have more experience in production and marketing will be better able to meet the quality expectations of direct market consumers and earn a higher profit. As seen from the literature review, this hypothesis was also supported by the findings of Corsi et al. (Reference Corsi, Borsotto, Borri and Strøm2009) and Park (Reference Park2015).
Both farm size and area under cherry orchard are hypothesized to be negatively correlated with the decision to participate direct marketing. These are consistent with the hypothesis put forward by Monson et al. (Reference Monson, Mainville and Kuminoff2008) and Wolanin (Reference Wolanin2013). Monson et al. (Reference Monson, Mainville and Kuminoff2008) put forward the hypothesis that since larger farmers can produce greater volumes, they may have an incentive to economize on their marketing costs by selling to buyers who can absorb a greater share of their production than direct market customers, who tend to make relatively small purchases. Additionally, Wolanin (Reference Wolanin2013) cites that smaller operations are more likely to rely on DTC outlets, perhaps because these operations may not be able to consistently produce quality fruits and vegetables or meet the high volume demands of intermediary or retail outlets. These hypotheses were also supported by the findings of Monson et al. (Reference Monson, Mainville and Kuminoff2008), Auld et al. (Reference Auld, Thilmany and Jones2009), Corsi et al. (Reference Corsi, Borsotto, Borri and Strøm2009), Timmons and Wang (Reference Timmons and Wang2010), Detre et al. (Reference Detre, Mark, Mishra and Adhikari2011), and Benedek et al. (Reference Benedek, Fertő, Baráth and Tóth2014) who concluded that adoption of a DMS has a negative relation to large farms.
Specialization in cherry production is hypothesized to be negatively correlated with the decision to participate in direct marketing. The fully-specialized farms which had low crop diversity did not generally want to take risk to sell directly to consumers due to lack of knowledge and experience. As seen from the literature review, this hypothesis was also supported by the findings of Aguglia et al. (Reference Aguglia, Francesco and Salvioni2009), Detre et al. (Reference Detre, Mark, Mishra and Adhikari2011), Park et al. (Reference Park, Mishra and Wozniak2011, Reference Park, Mishra and Wozniak2014), and Corsi et al. (Reference Corsi, Novelli and Pettenati2014) who concluded that as farms had become more specialized, they had a lower likelihood to market through DTC outlets.
Total annual farm income is hypothesized to be positively correlated with the decision to participate in direct marketing. This is also consistent with the hypothesis put forward by Monson et al. (Reference Monson, Mainville and Kuminoff2008) in their study. They put forward the hypothesis that as household income increases, reliance on direct marketing channels will increase. They also highlight that higher-income farmers may be wealthier as a result of their participation in direct marketing, or their income may serve as an enabling factor-income from farm and off-farm activities may be used to fund investment in high-value and direct marketing enterprises. As seen from the literature review, this hypothesis was also supported by the findings of Nyaupane and Gillespie (Reference Nyaupane and Gillespie2010) and Park et al. (Reference Park, Mishra and Wozniak2014) who concluded that farm income significantly affects farmer selection of marketing outlet.
Gross margin obtained per hectare for cherry production is hypothesized to have a positive effect on the cherry growers’ decisions to participate in direct marketing. As shown in Table 1, gross margin is the difference between gross income and variable costs. It represents how much each unit of an enterprise contributes toward fixed costs and profits, after the variable costs of production have been paid (Kay et al., Reference Kay, Edwards and Duffy2004). The above mentioned hypothesis was also supported by the finding of Park et al. (Reference Park, Mishra and Wozniak2014) who concluded that a significant level of management skill—more ways to control input costs—increases the likelihood that a farmer uses DTC outlets. Park et al. (Reference Park, Mishra and Wozniak2014) emphasized that management and marketing skills significantly affect DTC sales.
The growth in numbers of marketing channels used by farmers is hypothesized to be negatively correlated with the decision to participate in direct marketing. Farmers who used much fewer marketing channels for their products are more likely to use direct marketing channels, due to a lack of access to markets. As mentioned above, this hypothesis was supported by the finding of Park et al. (Reference Park, Mishra and Wozniak2014) who concluded that marketing skills significantly affect DTC sales.
There is a strict assumption that has to be made when using an ordinal regression model, the parallel lines assumption. That is to say, the regression coefficients are equal for all corresponding outcome categories (Minetos and Polyzos, Reference Minetos and Polyzos2007). Therefore, the test of parallel lines was designed to make a judgment concerning the adequacy of the model. The null hypothesis states that the corresponding regression coefficient is equal across all levels of the response variable. The alternative hypothesis states that the corresponding regression coefficients are different across all levels of response variable (Yay and Akinci, Reference Yay and Akinci2009).
Before proceeding to examine the individual coefficients, we looked at an overall test of the null hypothesis that the location coefficients for all of the variables in the model are 0. This can be based on the change in –2 log-likelihood when the variables are added to a model that contains only the intercept. The change in likelihood function has a chi-square distribution even when there are cells with small observed and predicted counts. If the chi-square has an observed significance level of less than 0.05, this means that we can reject the null hypothesis that the model without predictors is as good as the model with the predictors.
A standard statistical maneuver for testing whether a model fits is to compare observed and expected values. For this purpose, from the observed and expected frequencies, we computed the usual Pearson and deviance goodness-of-fit measures. Both of the goodness-of-fit statistics should be used only for models that have reasonably large expected values in each cell. If our model fits well, the observed and expected cell counts are similar, the value of each statistic is small and the observed significance level is large. We reject the null hypothesis that the model fits if the observed significance level for the goodness-of-fit statistics is small. Good models have large observed significance levels.
Cox and Snell's R 2, Nagelkerke's R 2 and McFadden's R 2, which are three commonly used statistics, were used to measure the strength of the association between the dependent variable and the predictor variables. These measures do not have the percent of variance explained and should not be reported in those terms. They can be taken as additional measures of model effect size, with higher values being better.
The exponentiated coefficients were used for the parameter's interpretation in the ordinary logistic regression model. The odds ratios can be obtained by exponentiation of the coefficients. P i /(1 − P i ) is referred to as the odds of an event occurring (Williams, Reference Williams2015). The effect of each variable on the odds comes from taking the antilog of the coefficients (exponentiated coefficients) (Pampel, Reference Pampel2000). In the case of an ordinal outcome with three or more categories, the odds ratio for the logit model represents the odds of the higher category, as compared with all lower categories combined (Heck et al., Reference Heck, Thomas and Tabata2013). Standard interpretation of the ordered logit coefficient is that for a one-unit increase in the predictor, the response variable level is expected to change by its respective regression coefficient in the ordered log-odds scale, while the other variables in the model are held constant (Chen et al., Reference Chen, Ender, Mitchell and Wells2003).
In order to measure the effects of explanatory variables on whether a farmer has a tendency toward direct marketing, the marginal effects were calculated. A marginal effect is the influence a one-unit change in an explanatory variable has on the probability of selecting a particular outcome, ceteris paribus. This holds for continuous variables only. For dummy (1, 0) variables, the marginal effects are the derivatives of the probabilities, given a change in the level of the dummy variable (Greene, Reference Greene2008).
Wald Ӽ2 statistics were used to test the significance of individual coefficients in the model and are calculated as follows (Bewick et al., Reference Bewick, Cheek and Ball2005):
$$= {\left( {\displaystyle{{{\rm coefficient}} \over {{\rm SE}\;{\rm coefficient}}}} \right)^2}$$
Each Wald statistic is compared with a Ӽ2 distribution with one degree of freedom. Wald statistics are easy to calculate, but their reliability is questionable, particularly for small samples.
Results and Discussion
The socio-economic characteristics of sample cherry farms
Table 2 summarizes the survey results on socio-economic characteristics of the cherry farms. The average age of the farm household operators was about 50 yr old and did not vary significantly across farm size. Operators had fairly long experience in cherry farming, and farmers had, on average, 28.64 yr of experience in cherry production. The operator's experience in cherry farming typically increased with farm size. The average number of years in school of the household operators in the sample was about 8 yr. The average years of schooling among smallholders was lower, at 7.41 yr, compared with 8.25 yr for household operators of medium- and large-scale farms. In terms of average farm area operated, the sampled large-scale farms operated on 6.02 ha, which is much higher than the landholdings of small-scale farms (1.29 ha) and medium-scale farms (2.85 ha). The average area of cherries grown was 1.5 ha and increased with farm size. It ranged from 0.59 ha in the case of small farms to 2.90 ha in the case of large farms.
Socio-economic characteristics of cherry farms.
1 The average exchange rates between Turkish Lira (TRY) and the US dollar (USD) for June 2012 is US$1 = 1.8161TL (CBRT, 2012).
The average farm household had an annual net income of nearly US$32,966. Large-scale farms, with US$56,628, had a higher average annual farm income per farm than other farm sizes; however, the average annual income of small-scale farms, with nearly US$19,634, was lower compared with other farm size categories. The average household size was about four. The household size of large farms, with 4.79 members, was more compared with other farm size categories.
Use of direct marketing options by cherry growers and options that they plan to use in the future
When interviewing farm operators about whether they have used direct marketing options, only two growers reported using direct marketing options. This does not mean that cherry growers do not tend toward direct marketing. One of the reasons for this was the lack of experience and knowledge in using direct marketing techniques among growers. The second reason appears to be that there are not any organizations that will help them meet their direct marketing goals and build direct marketing arrangements between themselves and their consumers.
In examining the possible direct marketing options that the cherry growers plan to use in the future in cases of selling directly to consumers, community-supported agriculture (CSA) emerged as having drawn the most attention (Table 3). CSA was also described as an important marketing strategy for small, mid-sized and large farms. Especially, medium-sized farms pay more attention to CSA compared with small and large farms. There were also statistically significant differences in favor of medium-sized farms.
Direct marketing options that cherry growers plan to use in the future.
$\bar x$
: the mean score of 5-point Likert scale (1 = absolutely not, 2 = preferably not, 3 = neutral, 4 = possibly, or 5 = definitively).
* Denotes significance at the 5% level.
On-farm sales (also known as on-farm stores) are considered to be the second most important option that the cherry growers plan to use in the future in cases of selling directly to consumers. The reason they are paying attention to this option may be due to the lack of cold storage to keep cherries fresh and the general absence of refrigerated transport.
The use of Internet or e-mail sales as a direct marketing option in the future was given low priority. Although 61.0% of the surveyed growers own a computer and 92% of growers have an Internet connection, the cherry growers were not interested in selling their products via the Internet due to a lack of marketing knowledge.
Due to a small number of farmers’ markets across the province and long distances to existing farmers’ markets, most of the growers surveyed believe that farmers’ markets are a less feasible option for direct marketing in the future. But farmers’ markets may be the best of the options if the numbers of farmers’ markets are spread across the country. For example, farmers’ markets also serve as a key direct marketing channel for small and mid-size US producers (Diamond and Soto, Reference Diamond and Soto2009), so with the recent growth in their numbers, one could perceive that the prevalence of direct sales may boost farm income. The presence, growth and new creation of farmers’ markets are some of the most apparent signals of consumer and producer interest in developing direct markets. The number of farmers’ markets in the USA has grown dramatically, increasing 226% from 1996 to 2012, with over 7800 farmers’ markets operating in the USA (Thilmany et al., Reference Thilmany, Tegegne and Hines2012).
The least preferred direct marketing options for cherry growers in planning strategies for future direct marketing initiatives are selling directly to neighbors and acquaintances, roadside stands, selling during cherry festivals or special events and peddling.
Analysis of factors affecting cherry growers’ decisions to participate in direct marketing
An ordered logistic regression model was used to identify the factors that influence cherry growers’ decisions to participate in direct marketing. Firstly, we evaluated the appropriateness of this assumption through the ‘test of parallel lines’ to check whether the model meets the parallel lines assumption or not. Table 4 given below shows that the assumption is met, as the test shows a level of nonsignificance (0.271 > 0.05).
Test of parallel lines.
The null hypothesis states that the location parameters (slope coefficients) are the same across response categories.
From Table 5, we see that the difference between the two log-likelihoods—the chi square—has an observed significance level of less (P = 0.042) than 0.05. This means that we can reject the null hypothesis that the model without predictors is as good as the model with the predictors.
Model-fitting information.
Table 6 gives the Pearson and deviance goodness-of-fit measures of the model. We see that the goodness-of-fit measures have large observed significance levels, so it appears that the model fits.
Goodness-of-fit statistics.
* P values are greater than 0.05.
Table 7 shows the pseudo R 2 values for the strength of association between the variables. The values of the pseudo R 2 show a moderate size effect.
Pseudo R 2.
The estimated coefficients for the factors affecting the cherry growers’ tendency toward participation in direct marketing, along with their Standard errors, Wald test values, P-values, odds ratios and marginal effects, are presented in Table 8. Results in Table 8 show that the coefficients of age of the grower, the grower's level of education, total annual farm income, farm size and the number of marketing channels used by growers are not statistically significant. Finally, parameter estimates for these variables have no significant impact on the cherry growers’ tendency toward participation in direct marketing.
Ordered logit regression model estimates evaluating factors affecting cherry growers’ decisions to participate in direct marketing.
* and ** denotes significance at the 5 and 10% levels, respectively.
1 Marginal effects were calculated based on category 3 which denote the strongest tendency of direct marketing.
On the other hand, the results show that the cherry farming experience of the farmers, the size of the cherry orchard, the level of specialization in cherry production and the gross margin per hectares for cherry production have a statistically significant impact on the cherry growers’ tendency toward participation in direct marketing.
The category that is statistically significant in terms of years of experience in cherry growing is the growers who have experience of between 11 and 20 yr. This category was compared with a reference category. The last category, or category 3, was used as the reference group, which includes the growers with more than 20 yr experience in cherry growing. The exponentiated coefficients, or odds ratios, reported in Table 8 show that the growers having experience of between 11 and 20 yr tend to use direct marketing about 0.4 times less than the farmers with more than 20 yr experience. In other words, the growers with more than 20 yr experience tend to use direct marketing 2.53 (1/0.396) times more than growers having experience of between 11 and 20 yr. Based on an assessment of marginal effect, the growers having experience of between 11 and 20 yr tend to use direct marketing 9.4% less than the growers with more than 20 yr experience. The results suggest that the more experienced growers in cherry farming are more likely to sell their product directly to consumers. This is because growers who have more experience in cherry marketing have a greater need to seek alternative marketing options compared with less experienced farmers. They have been forced to accept low prices when selling their produce to indirect markets for a long time. Therefore, it is highly recommended to work with more experienced cherry growers in projects focused on the development of direct marketing in terms of achieving the desired results.
This result supports the hypothesis that as producers’ experience in agriculture increases, reliance on direct marketing channels will increase (Monson et al., Reference Monson, Mainville and Kuminoff2008). As noted by Monson et al. (Reference Monson, Mainville and Kuminoff2008), producers who have more experience in production and marketing will be better able to meet the quality expectations of direct market consumers and earn a higher profit. It is also possible that these producers will have more established ties and extended relationship networks within the community, which will enhance their ability to market to locals. Considering the fact that farmers typically gain skills through work experience, this result is consistent with the finding of Corsi et al. (Reference Corsi, Borsotto, Borri and Strøm2009) who concluded that more skilled farmers were less likely to choose traditional marketing chains and more likely to engage in the new marketing chains. This result also supports the finding of Park et al. (Reference Park, Mishra and Wozniak2014) who concluded that increasing management skills—more ways to control input costs—increased the likelihood that a farmer used DTC marketing outlets. However, Uva (Reference Uva2002) warned that a DMS required a set of skills different from those for agricultural operations.
While examining the impact of the size of total acreage farmed that is dedicated to cherry production on the cherry growers’ tendency toward participation in direct marketing, the total acreage devoted to cherry production was divided into three categories. When considering the statistical significance of the parameters for the categories, only the coefficient on cherry growers having cherry orchards between 1 and 2 ha was statistically significant. The farms that have more than 2 ha of cherry orchards were taken as the reference category. Accordingly, the farms between 1 and 2 ha tend to use direct marketing nearly 3 times (2.975) more than farms of over 2 ha. Based on an assessment of marginal effect, the growers who have cherry orchards between 1 and 2 ha tend to use direct marketing options 12.8% more than growers with cherry orchards of over 2 ha. This result supports generally the hypothesis that as farm size increases, producers become less reliant on direct marketing channels to sell their output. As noted by Monson et al. (Reference Monson, Mainville and Kuminoff2008), there are several justifications for this hypothesis. Firstly, larger farms are in a better position to overcome the barriers to entry to other potentially lucrative markets, such as supermarkets. Additionally, since larger farmers can produce greater volumes, they may have an incentive to economize on their marketing costs by selling to buyers who can absorb a greater share of their production than direct market customers, who tend to make relatively small purchases. This finding is generally consistent with findings from previous studies (Monson et al., Reference Monson, Mainville and Kuminoff2008; Auld et al., Reference Auld, Thilmany and Jones2009; Corsi et al., Reference Corsi, Borsotto, Borri and Strøm2009; Timmons and Wang, Reference Timmons and Wang2010; Detre et al., Reference Detre, Mark, Mishra and Adhikari2011; Benedek et al., Reference Benedek, Fertő, Baráth and Tóth2014), which suggest that large farms rather choose traditional chains rather than the direct and the short chains. These previous studies have demonstrated that adoption of a DMS has a negative relation to large farms.
The study also showed whether specialization in cherry production may impact growers’ tendency to participate in direct marketing. Farms that are fully specialized in cherry production were taken as the reference category. We found a statistically significant result in terms of nonspecialized farms in cherry production. Accordingly, farms that are not fully specialized in cherry production tend to use direct marketing 3 times (3.027) more than fully specialized farms. Based on an assessment of marginal effect, farms that are not fully specialized in cherry production tend to use direct marketing options 12.3% more than fully specialized farms. A significant proportion (more than 80%) of the total gross production value is derived from cherry cultivation in the fully specialized farms. On the other hand, the value of production for cherries produced on the not fully specialized farms accounts for a small proportion of the total gross production value. These farms mainly produce fresh peaches, compared with other products. It is possible to say that these farmers shift toward semi-specialized farming.
This result indicates that semi-specialized farms have a higher tendency to use direct marketing techniques than fully specialized farms. The major reason for this situation was that semi-specialized farms had a higher level of crop diversification than fully specialized farms. The fully-specialized farms, which had low crop diversity did not generally want to take risk to sell directly to consumers due to lack of knowledge and experience. This result reveals that direct marketing options are more appropriate for semi-specialized farms. This result supports findings by Aguglia et al. (Reference Aguglia, Francesco and Salvioni2009), Detre et al. (Reference Detre, Mark, Mishra and Adhikari2011), Park et al. (Reference Park, Mishra and Wozniak2011), and Corsi et al. (Reference Corsi, Novelli and Pettenati2014), who concluded that as farms had become more specialized, they had a lower likelihood to market through DTC outlets. This result is also consistent with the findings of Park et al. (Reference Park, Mishra and Wozniak2014) who concluded that farmers engaged in direct sales only tend to market a greater diversity of crops and have a larger variation in the earnings shares of those crops compared with farmers with no direct sales.
While examining the impact of gross margins obtained per hectare for cherries on cherry growers’ tendency toward participation in direct marketing, the gross margins obtained per hectare were divided into three categories. When considering the statistical significance of the parameters for the categories, only the coefficient on cherry farms that achieve a gross margin between US$2753 and 5506 ha−1 for cherries was statistically significant. The farms that achieved more than US$5506 ha−1 were taken as the reference category. Accordingly, farms that achieve a gross margin between US$2753 and 5506 ha−1 tend to use direct marketing 0.3 times (0.321) less than farms that achieve more than US$5506 gross margin per hectare. In other words, farms that achieve more than US$5506 gross margin per hectare tend to use direct marketing over 3 times (1/0.321) more than farms that achieve a gross margin between US$2753 and 5506 ha−1. Based on an assessment of marginal effect, farms that achieve a gross margin between US$2753 and 5506 ha−1 tend to use direct marketing options 10.4% less than farms that achieve more than US$5506 gross margin per hectare.
There are a number of possible explanations for this result. One possible explanation is that this may be due to the farm management and marketing skills gathered over the years. As noted by Monson et al. (Reference Monson, Mainville and Kuminoff2008), as producers gain experience in production and marketing, the greater the reduction in costs which will result in higher profits. Monson et al. (Reference Monson, Mainville and Kuminoff2008) also cite as producers’ experience in agriculture increases, reliance on direct marketing channels will increase. Hardesty and Leff (Reference Hardesty and Leff2010) warned that farmers must manage their marketing costs as well as their production costs in order to get success in direct marketing. Park et al. (Reference Park, Mishra and Wozniak2014) emphasized that management and marketing skills significantly affect DTC sales. Park et al. (Reference Park, Mishra and Wozniak2014) determined that a significant level of management skill—more ways to control input costs—increases the likelihood that a farmer uses DTC outlets.
Conclusions
The objective of this study was to explore the main factors that drive the decisions of farmers to sell their products directly to consumers through farm direct marketing channels. A case study on cherry growers on the subject of direct marketing, which is one of the alternative marketing options in agricultural products marketing for farmers, was examined in this study. The major findings of this study are briefly outlined below. In addition, further suggestions were put forward on how to improve the use of DMSs by farmers in Turkey.
While examining the cherry farmers who were interviewed for possible marketing strategies in cases of selling directly to consumers, community-supported agriculture emerged as having drawn the most attention.
An ordinal logistic regression analysis model was used to analyze the effects of agricultural businesses and demographic features on the growers’ tendency to choose direct marketing channels in cherry selling. According to these model results, the cherry farming experience of the growers, the size of the cherry orchard, the level of specialization in cherry production and the gross margin per hectare for cherry production have a statistically significant impact on the growers’ tendency to choose direct marketing channels in cherry selling. In particular, the growers whose experience is more than 20 yr, the farms that are semi-specialized, the farms providing a gross margin of more than US$5506 ha−1 and the farms having a cherry orchard between 1 and 2 ha in size were determined as having more tendency toward direct marketing. The results obtained allow us to conclude that the farmers who have more experience of growing cherries, operate a medium-sized farm growing cherries, are semi-specialized in cherry production, and have a higher gross margin per hectare for cherries are more likely to sell their produce through DTC outlets. The results from this analysis may help agricultural extension agents and other institutions in Turkey in facilitating the adoption of DMSs. As aforementioned, even though there are some studies done to explore factors that may be related to a farmer's decision to participate in direct marketing, the effects of these factors may vary on the bases of ‘region-case’ and ‘crop-case’. Thus, further research should be assessed on the bases of both cases.
According to the interviewed cherry growers, the most important factor limiting their participation in direct marketing is that there are not any organizations that will help them meet their direct marketing goals and build direct marketing arrangements between themselves and their consumers. The other important factors limiting grower participation in making the decision to participate in direct marketing are the lack of cold storage to keep cherries fresh, the challenge of reaching consumers, the amount of time required to operate a direct market, the general absence of refrigerated transport, having a concern whether all the cherries grown in farms are marketed to consumers through direct marketing channels and the lack of experience and knowledge in using direct marketing techniques among growers.
Considering all the factors that can limit grower participation in making the decision to participate in direct marketing, it is important that various organizations, such as producers’ associations, cooperatives, local administrations, voluntary consumer groups around business and family, and civil society organizations, take an active role as facilitators in organizing sales directly to consumers. Especially, local administrations can play a key role for faster dissemination and adoption of direct farm marketing initiatives, such as CSA programs, pick-your-own activities, farmers’ markets and agri-tourism. However, local administrations should work with civil society organizations to provide an effective functional link between farmers and consumers. These organizations should teach farmers how to use direct marketing to reach target audiences. In addition to training farmers on direct farm marketing, these organizations should often provide technical assistance, networking, marketing opportunities and other services to support direct marketing. The findings from this study can guide policy makers in improving direct marketing plans and programs on the marketing of fresh fruits.
Acknowledgements
Funding for this research was provided by the Ege University Scientific Research Fund under grant number 2011-ZRF-051. The author would like to thank Ege University Research Fund for its financial support. The author also wishes to thank the cherry growers who contributed to the farm survey.
