Consumption Preference and Willingness to Pay for Locally Produced Rice in Namibia: Implications for Upscaling Production

This study was conducted in two towns and one settlement in the Omusati region of northern Namibia: Outapi, Oshikuku, and Ogongo. The target population for this research comprised residents of the area with purchasing power.

This study employed a mixed-methods approach, integrating qualitative and quantitative data collection techniques. Experimental auctioning, combined with sensory evaluation, was used to elicit preferences and WTP. The experimental sessions were conducted over two weeks, during which questionnaires were administered. These questionnaires collected qualitative data (e.g., ranking likelihood using a hedonic scale), quantitative (e.g., bids), and sociodemographic data.

Following the methodology of Demont et al. (2012), this study focused on two locally grown rice varieties, Irriga and Supa. Additionally, Tastic Rice, a type already available on the Namibian market, was used as a reference point for comparison with the two local varieties.

The second-price auction was selected for its ability to ensure that participants revealed their true valuation of goods through their bids. This approach was employed alongside the endow-and-upgrade method, where participants were endowed with a certain quantity of the benchmark rice and asked to indicate the premium price they were willing to pay to upgrade to one of the two local varieties.

All rice types were presented to participants as specimens. The details of the two local varieties (such as their names and origins) were not disclosed. Instead, the varieties were distinguished based solely on their attributes. Participants assessed the rice attributes visually (e.g., purity level and uniformity) and sensorially (e.g., tasting, aroma, swelling capacity). Using a 9-point hedonic scale, participants rated how much they liked each rice variety’s attributes.

A quantitative sampling technique, specifically simple random sampling, was employed to minimise personal bias. Samples were drawn from two towns and one settlement in the Omusati region of northern Namibia: Outapi, Oshikuku, and Ogongo.

Since the size of the study population was unknown, the Cochrane method was used to calculate the optimum sample size for use in this study.

The Cochran formula is expressed as: 1no=Z2pqe2{n_{\rm{o}}} = {{{Z^2}pq} \over {{e^2}}}

Where: n_o – is the sample size, e – is the desired level of precision, which was set to 0.08 for a confidence level of 92%, z – is the z-value corresponding to the given confidence level, p – is the estimated proportion of the population with the given attribute; if the population proportion is unknown, we assume it to be 50% for estimation purposes, q – is 1 – p.

Using this formula, the target sample size was calculated to be 118 respondents. However, the final sample size for the study consisted of 105 respondents.

The questionnaire was designed to collect both qualitative and quantitative data, including participant demographics. Qualitative data were recorded using a 9-point hedonic scale, while quantitative data were obtained from the WTP values derived from auction bids.

The data collected for this study were analysed using the Statistical Package for the Social Sciences (SPSS). Kendall’s Coefficient of Concordance was employed to rank the quality attributes that consumers considered important when buying rice. This statistical measure helps identify and rank a given set of attributes from the most influential to the least influential and assesses the degree of agreement or concordance among respondents regarding their perceived importance (Al-Hassan et al., 2009).

Kendall’s Coefficient of Concordance was used to rank the identified quality attributes, with ranks assigned from 1 to n, where 1 represents the most important attribute and n represents the least important. The model is specified as follows: W=∑ T2−12[ ∑ T2−(∑ T)2n ]nm2(n2−1)W = \sum {{T^2}} - {{12\left[ {\sum {{T^2}} - {{{{\left( {\sum T } \right)}^2}} \over n}} \right]} \over {n{m^2}\left( {{n^2} - 1} \right)}}

Where: T – sum of ranks for attributes being ranked, m – number of respondents (consumers), n – number of attributes being ranked, W – coefficient of concordance.

The multinomial logistic regression is specified as: 4Ln(P/1−P)=βo+β1X1+β2X2+β3X3++β4X4+β5X5+β6X6+β7X7+β8X8+β9X9+β10X10+β11X11+e\matrix{ {{\mathop{\rm Ln}\nolimits} (P/1 - P) = {\beta _{\rm{o}}} + {\beta _1}{X_1} + {\beta _2}{X_2} + {\beta _3}{X_3} + } \cr { + {\beta _4}{X_4} + {\beta _5}{X_5} + {\beta _6}{X_6} + {\beta _7}{X_7} + {\beta _8}{X_8} + {\beta _9}{X_9} + {\beta _{10}}{X_{10}} + {\beta _{11}}{X_{11}} + e} \cr }

Where:

• Dependent variable:

  Y – rice A, rice B, rice C.

• The independent variables were as follows:

  X₁ – smell, X₂ – texture, X₃ – taste, X₄ – shape, X₅ – stickiness, X₆ – swelling capacity, X₇ – packaging, X₈ – brand, X₉ – label, X₁₀ – price, X₁₁ – promotion

• Other terms:

  e – confounding variance, Ln – Natural logarithm, P – probability of consumer WTP for a specific rice based on an attribute, β_o – regression constant or intercept, β_1,2,3…n – independent variable coefficients

A total of 105 participants were included in this study, representing the total number of participants in the five experimental sessions conducted in the Omusati region. Sociodemographic characteristics of rice consumers are important for conducting marketing studies, such as WTP, and for assessing consumers’ consumption patterns (Campiche et al., 2004). A summary of the demographic characteristics of the participants, including age, sex, level of education, and employment status, is shown in Table 2.

In this study, there were more female participants (66.7%) than male participants (33.3%). The majority (56.2%) of the sample were aged 25 years or younger, followed by 25.0% in the 26 to 35 years age range. Among the participants, 15.2% were married, 83.8% were single, and 1% were divorced. The majority (63.8%) of consumers were students, followed by those (22.9%) with a source of income from either full-time (20%) or part-time (2.9%) work. Table 2 also shows that a minority (3.8%) of consumers earned a monthly income greater than N$17,500 and another 3.8% earned between N$10,001 and N$17,500. Most consumers (83.8%) earned less than N$2,500, followed by 8.6% who earned between N$2,501 and N$10,000. More than half (52.4%) of the respondents were unaware of any locally grown rice, while 47.6% indicated that they were aware of some local rice varieties grown within the country.

The mean preferences for the rice types are presented in Table 3, showing that consumers had a high preference for Tastic rice (mean = 6.80), followed by Irriga (mean = 5.54), and Supa (mean = 5.04), with a statistically significant difference among the preference means (p < 0.001).

The mean preferences for Tastic, Irriga, and Supa are shown in Figure 4. The results indicated that the mean for Tastic rice was higher than that for Irriga and Supa. However, Supa rice was the least preferred due to its stickiness and fragrance/aroma. Most respondents also reported that they did not know how to cook Supa rice using methods for other parboiled rice, such as boiling the rice and draining the water. Supa rice requires minimal water that must be completely absorbed during cooking, necessitating careful measurements of water and, in some cases, special rice cooking pots.

Fig. 4.

Mean preference ranking comparison of the various uncooked rice types

Source: survey data.

The descriptive statistics for the rice attributes rated by consumers for the three rice varieties are summarised in Table 4. The most preferred uncooked attribute was the grain size of Tastic rice (7.10), followed by Irriga (5.59), and Supa (5.45). Other preferred attributes of uncooked Tastic rice included texture (6.98) and shape of the kernels (6.76). For cooked rice, Irriga was preferred for its aroma (6.61) and colour (7.11). All attributes for both uncooked and cooked rice were statistically different between rice types (p < 0.05). Supa rice was the least preferred in both categories.

Kendall’s coefficient of concordance analysis for both extrinsic and intrinsic quality attributes that consumers considered important when buying rice was conducted (Table 5). Kendall’s W measures the level of agreement in rankings among respondents, with a lower coefficient (closer to zero), indicating a low level of agreement, and a higher coefficient (closer to one), indicating a high level of agreement (Danso-Abbeam et al., 2014). A high mean indicates greater importance, whereas a low mean indicates lesser importance (Al-Hassan et al., 2009). The Kendall’s W for this study was 0.061, or 6.1% agreement, despite being statistically significant (p < 0.001).

The results showed that taste was ranked first with the highest mean rank (7.5), followed by price (7.46), and swelling capacity (6.9). The least important attributes were smell/aroma/fragrance, texture, and packaging, which were ranked 11^th, 10^th, and 9^th, respectively.

Table 6 presents a descriptive summary of the means and standard deviations of WTP for local and imported rice. The results show that respondents were ready to pay more (N$28.68 ±24.42) per kilogram for imported rice (Tastic) compared to locally grown rice (N$25.98 ±18.06 for Irriga and N$20.05 ±10.11 for Supa). However, after sensory evaluation of cooked rice, consumers’ WTP decreased for all rice types, but the pattern and order of prices remained the same as uncooked rice. Additionally, there was a statistically significant difference between the prices of the three types of uncooked rice (p = 0.0054) and cooked rice (p = 0.00037).

The post hoc test revealed statistically significant differences in WTP between Tastic (A) and Supa (C) for uncooked rice (p=0.0043). For cooked rice, there were statistically significant differences in WTP between Tastic (A) and Supa (C) (p = 0.0004), and between Irriga (B) and Supa (C) (p = 0.0108) (Table 7).

A multinomial logistic regression analysis was conducted to determine the effect of extrinsic and intrinsic quality attributes on WTP for the different rice types, using Tastic rice as the reference category (Table 8). The model showed a good fit, although it was not statistically significant. The explanatory variables accounted for less than 30% of the variance, indicating that 70% of the variance in the dependent outcome variable was explained by other variables not included in the model.

Although the model was not significant, some of the predictor variables for the different rice types showed significant values (Table 9). For Irriga, promotion was the only significant attribute at the 10% confidence level. For Supa, swelling capacity, price, and promotion were the significant variables. Supa swelling capacity and promotion were both significant at the 10% confidence level, while price was significant at the 5% confidence level.

Owing to the demographic characteristics of the respondents in this study, a greater proportion of participants were female (66.7%) compared to male participants (33.3%). This is significant, as a higher proportion (75.7%) of female respondents (out of the total number of female respondents) stated that they consumed rice often, compared to 65% of male respondents, who reported frequent rice consumption. Only 3.8% of the consumers earned a monthly income of more than N$17,500, and another 3.8% earned within the range of N$10,001 to N$17,500. The largest proportion (83.8%) of consumers earned less than N$2,500, followed by those earning between N$2,501 and N$10,000 (8.6%). This is likely due to the presence of a University of Namibia campus within the study area, which leads to a high student participation rate. However, Ashraf and Merunka (2017a) found that students represent a userfriendly population that exhibits convenience. They also argued that rejecting data collected from students could hinder science development.

The findings of this study reveal that consumers prefer imported rice (Tastic). This preference may stem from participants’ unfamiliarity with the two local rice varieties, particularly the unique attributes of stickiness and aroma that these varieties exhibited. Familiarity with food and cooking methods influences consumer preferences, as individuals tend to favour foods they recognize and know how to prepare. As familiarity increases, so does preference and WTP (Frez-Muñoz et al., 2024; Goulart et al., 2021; Manohar et al., 2021). Moreover, Supa rice, known for its stickiness and strong aroma, was unfamiliar to many respondents, making it less desirable. This finding aligns with Tuorila and Hartmann (2020), who suggested that foods perceived as unnatural are less desirable.

Consumers favoured the grain size and swelling capacity of Tastic rice over Irriga and Supa rice. The most disliked attributes of the local rice varieties were stickiness and aroma (see Table 4). A majority of respondents expressed a dislike for the stickiness of Supa rice, which many likened to pap, resulting in Supa being ranked lowest (4.81) compared to Irriga (6.12) and Tastic rice (6.76). This suggests that most respondents prefer less sticky rice, which may have implications for the development of local rice varieties—unless significant efforts are made to promote the benefits of sticky and aromatic rice and teach consumers how to prepare it effectively. Promotion plays a key role in introducing new products to the market and increasing consumer familiarity (Antriyandarti et al., 2023). Supa, in particular, differs from the preparation methods for the other two rice types, as it must be cooked at low heat until all the water is absorbed, posing challenges in terms of managing heat and water levels. Many respondents reported cooking Supa unsuccessfully, resulting in it becoming soggy and unappetizing, further explaining their dislike for it.

Imported rice, which shares characteristics with Supa (such as stickiness and aroma) is often quite expensive and out of reach of the average household income, possibly explaining consumers’ unfamiliarity with the properties of Supa. Similar types of rice such as Jasmine and Basmati, are typically aromatic and costly, making them accessible primarily to elites or high-income individuals who may have travelled and are familiar with these varieties.

The tastes of Irriga and Supa were also found to be less appealing to the consumers. Most respondents noted that these local varieties were not as flavourful as the imported rice used in this study. As shown in Table 4, the taste ratings for Irriga and Supa were lower (6.61 & 5.08, respectively) compared to Tastic rice (6.85). Additionally, the two local varieties were noted to have a stronger smell than the imported rice, further driving consumers’ preference for the latter. This suggests that consumers favour rice with a mild fragrance rather than one with a strong aroma, possibly due to the prevalence of imported rice types in the Namibian market, which have a milder scent.

However, Irriga’s whiteness – its pure, white colour – was the most preferred attribute among the three rice types. Cooked Irriga is visually appealing to consumers due to its pristine colour. These preferred attributes of locally grown rice should be emphasized to improve consumer preference. Additionally, these findings have breeding implications, as consumers seem to favour rice with large grains, good swelling capacity, and a white appearance accompanied by a mild fragrance.

A similar study conducted in Haiti by Pavilus (2018) found contrasting results, with consumers in Haiti preferring local rice over imported rice. However, some attributes of the local rice preferred by Haitians were similar to those identified by participants in this study, such as a preference for non-sticky rice with a mild aroma. This suggests that the quality attributes of rice play a crucial role in shaping consumer preferences for both imported and locally produced rice varieties.

The attributes that are considered important by consumers when buying rice include price, taste, promotion, swelling capacity, stickiness, and brand. These findings are consistent with the results of other studies, such as Sampson (2013) and Laizer et al. (2018), which also indicate that price and taste are among the most important attributes influencing consumers’ preferences of rice. According to Laizer et al. (2018), aroma is the most important rice attribute characterising consumer preferences. However, in this study, aroma was ranked as the second-least important attribute. This difference may be due to the Namibian market, which is characterised by an abundance of non-fragrant rice types. Additionally, consumer behaviour and preferences are influenced by sensory attributes, marketing factors, and psychological factors (Font-i-Furnols and Guerrero, 2014). These factors, which fall outside of the scope of this study, could have played a significant role in its findings.

Kendall’s coefficient of concordance (W) revealed that 6.1% of respondents provided similar rankings based on their preferences for different rice attributes. This suggests that their rankings varied in terms of how they prioritised the attributes. The model assumes a null hypothesis where the rankings disagree. Given that the analysis is significant at the 99.9% confidence level, we reject the null hypothesis and accept the alternative hypothesis that the rankings agree. Therefore, we conclude that consumers generally agree that the most important rice attribute is price, followed by taste, promotion, swelling capacity, stickiness, brand, and packaging. The shapes of the kernels, labels, smells, and textures were ranked as the least important characteristics.

However, the low percentage of agreement (6.1%) shown by Kendall’s coefficient is due to the variation in the data. This means that very few respondents gave the same responses. The variation in the rankings of importance contributed to the low percentage of agreement, as indicated by Kendall’s coefficient of concordance.

The study found that consumers were not willing to pay more for local rice varieties. Since this study was conducted near a rice production area, it was anticipated that respondents might be influenced by a social norm that devalues imported rice (Demont et al., 2012). However, this outcome was not surprising, as previous studies have shown that, in some cases, consumers consistently prefer imported rice to local varieties and are even willing to pay more for imported rice, despite local varieties often offering superior quality attributes (Obih and Baiyegunhi, 2017).

After a sensory evaluation of cooked rice, the average willingness to pay (WTP) decreased for all three rice types. Supa exhibited the most notable decrease in WTP, dropping by 8.4%, followed by Tastic with a 3.7% decrease, and Irriga with a smaller decline of 2%. This pattern suggests that respondents were less satisfied with the sensory attributes of cooked Supa rice compared to the other varieties. These results are consistent with Demont et al. (2012), who found that WTP typically decreases following a sensory evaluation of cooked rice attributes.

Al-Hassan et al. (2009) found that consumers are generally willing to pay more for aromatic rice. However, in this study, consumers did not show a willingness to pay more for rice varieties with familiar aromas. This indicates that consumers were unwilling to pay more for the two local rice varieties, possibly because they were not familiar with the specific quality attributes of these varieties.

This study also found that women in the study area dominate rice purchasing and consumption. Therefore, when improving the rice production system, researchers should consider women’s preferences and characteristics. This finding supports the idea that food purchasing decisions are primarily made by women, rather than men (Loginova and Mann, 2024).

The study revealed that knowledge and consumption of local rice are very limited in the area. Many consumers were more familiar with imported rice than with local rice, which influenced their preferences and WTP for imported rice. This could impact local rice production in the country if the population continues to prefer imported rice. The study concluded that rice products should be promoted, advertised, and packaged in a way that appeals to new buyers. This includes providing information about cooking, preparation, and incorporation into familiar dishes.

However, when asked if they would be willing to buy local rice if production increased, most respondents answered yes. This is a positive sign for potential future demand for local rice, as well as opportunities for the development of the Namibian rice industry. Additionally, land availability for rice production presents a significant opportunity for expansion.

The study also identified several reasons why consumers do not buy locally produced rice. These reasons include limited access, lack of awareness, lack of branding, no cooking instructions, and poor quality of locally produced rice. However, individuals who consume local rice do so because they consider it to have good nutritional quality and find it less expensive.

Consumers’ rice choices were influenced by certain quality characteristics, and they were willing to pay more for rice with desirable attributes. However, the study’s findings also showed that quality attributes and socio-demographic characteristics of the respondents do significantly affect WTP. This study assumed that the insignificant influence may have been due to the homogeneous nature of the data used. Therefore, further research is needed to obtain conclusive results on the determinants of WTP, particularly research that examines psychological, sensory, and marketing factors across a wider and more representative consumer sample. Given that rice grows well in areas prone to flooding, it could be a key crop for food security in north-central Namibia, where floods and droughts are common and climate change adaptation is necessary. This study provides valuable insights and contributes to the literature on market analyses, consumer behaviour, and sensory evaluation, climate change adaptation, food security, and local agricultural development.