![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
ABSTRACT
The identification of the market potentials of organic products is important in the drive towards a sustainable agricultural development in sub-Saharan Africa (SSA). However, available evidence shows that valuing attributes of credence goods (such as organic products) while using stated preference methods faces additional obstacles compared to other normal goods. In this study, we examine consumers’ preferences and willingness-to-pay (WTP) for health and environmental attributes of organic products in Nigeria. We employ an approach that allows us to adequately capture the value of organic products by linking part of the heterogeneity across respondents to differences in scale, while making use of indicators of survey engagement, without risks of endogeneity bias and measurement error that arise from the deterministic methods. The empirical results show that market for organic products exists in Nigeria, with reduction in pesticide residues attribute attracting the highest value, followed by the certification programme. Furthermore, we observe that increases in the latent engagement variable lead to a greater probability of agreement with statements relating to survey understanding and realism, and hence more substantive output.
Notes
1 Currently, of the 11 987 hectares of land under OA, less than 60 hectares are recorded as fully certified organic farms and virtually all the organic products are for export.
2 The underlying perception is that variations in scale across respondents constitutes a significant share of the heterogeneity in random coefficients models, rather than differences in sensitivities.
3 They consider individual differences in the ability to deal with complexity, arising due to variations in demographic variables, such as literacy, age, experience and cognitive ability, among others.
4 Ready, Champ, and Lawton (Citation2010) reveals that ex post approach is highly complex in CEs having more than two options per choice scenario, which is the case in our study.
5 We are unable to conduct a nonhypothetical stated preference approach (i.e., the experimental auction method) due to the fact that the organic product concepts tested in this study are yet to be available on the market. More so, auction methods are more expensive and time-consuming to implement as subjects have to be paid a participation fee and actual transactions have to be made during the experiment. Studies have shown that ideally, an analyst must possess all the product profiles presented in the choice sets in order to properly execute an experimental auction, given that it involves the exchange of real money for actual products (e.g., Harrison Citation2006, Wu et al. Citation2015).
6 Exploratory factor analysis was employed to test the reliability and internal consistency of the indicators. The value of Cronbach’s alpha (0.733) confirms reliability of using these indicators as a common construct.
7 The final design was generated using the Ngene software (version 1.0) and we accounted for uncertainty of priors by employing normally distributed Bayesian priors. The final design with the lowest Bayesian D-error (0.2534) was attribute-level balanced.
8 Tomato production plays an important role in enhancing food security in Nigeria, as it provides food and raw materials for industries, income from sales, and employment for smallholder households in urban and peri-urban areas. The demand for tomato is universal in the country, it serves as an excellent source of good amount of vitamin C and beta-carotene, and also there are no cultural/religious barriers against it. Tomato makes up about 18% of the average daily consumption of vegetables in Nigerian homes. Furthermore, Nigeria is ranked the largest producer of tomato in SSA and thirtieth largest in the world with an annual total area of one million hectares used for tomato cultivation and about 1.701 million tonnes of tomatoes produced annually, at an average of 25–30 tonnes per hectare (FAO Citation2010). However, tomato being a perishable product remains susceptible to location- and cultivar-specific pests and diseases. Thus, as farmers attempt to meet growing demand and are faced with strong pest pressure, they increasingly rely on synthetic pesticides to reduce the risk of harvest and income loss (e.g., Lund et al. Citation2010).
9 In principle, organic certification can improve producers’ environmental performance, even in countries where state regulation is weak.
10 A sample of 2700 observations from 300 respondents, each in the HP and CT treatments were used for the analysis.
11 Increasing the draw to 500 and 1000, did not have marked impact on our results.
12 In estimating the models, we observe that the medium level of the attributes were not statistically significant from zero, thus for the reason of parsimony, the medium and base levels were effectively collapsed to form a single base level (e.g., Collins, Rose, and Hess Citation2012).
13 For example, the Cholesky terms, and
give the two components of the Cholesky matrix relating to the pesticide coefficient, the first being off-diagonal, the second being the diagonal element, while e.g.,
gives the mean distribution for the pesticide coefficient.
14 We also estimated the models with normally distributed coefficients, but these were found to have lower goodness of fit.
15 However, in our final model specification, no significant alternative specific constants were recovered, and we thus limited ourselves to the effects of five explanatory variables.
16 For the negative price attribute, given that the log-normal distribution produces positive parameter value which may be contrary to a priori expectation, we follow the literature and reverse the sign by defining the negative of the attribute prior to model estimation.
17 Here, we use the same draws as those used in estimation, and incorporate the socio-demographic shifts applicable to each respondent (e.g., Hess and Stathopoulos 2013).
18 We take advantage of the properties of maximum likelihood and simulate multiple datasets by drawing 10 000 observations from a multivariate normal distribution parameterized by the means and covariances that arise from the estimations.
19 Our survey data bordered mainly on consumers’ stated choices and follow-up questions on their engagement, we did not capture information on organic production factors and input costs. Thus, the analysis of production data for organic tomato is beyond the scope of our study. Nevertheless, the WTP for organic products attributes found in this research is clearly within the range of price premiums identified by other studies. Although evidence from developing countries is limited, the review by Yirdidoe, Bonti-Ankomah, and Martin (Citation2005) suggests an average WTP premium for organic certified goods of about 30%, and Coulibaly et al. (Citation2011) on their study of households in urban Ghana and Benin, calculate a premium for organic certification of 57–66% for cabbage and 50–56% for tomatoes.
20 This approach compares the differences between every combination of data points in the empirical distributions that arise from the bootstrapping procedure. For iterations of the bootstrapping procedure, the Poe, Giraud, and Loomis (Citation2005) method considers differences. Thus, for a bootstrap procedure with 10 000 iterations, this would imply 10 0002 = 100 000 000 differences. To make these computations tractable, we reduced the number of data points from 10 000 down to 1000 for the complete combinatorial test.