This study was conducted in the Federal Capital Territory (FCT) located in North Central Nigeria (Figure 2). It lies within latitudes 7 25’ and 9 25’N and longitude 5 45’ and 7 39’E. It is geographically located in the savannah vegetation and at the centre of the country with a landmass of 7315 km². Federal Capital Territory is characterised by alternating dry and wet seasons with a mean annual rainfall that varies from 1100 mm to 1600 mm and temperatures of between 12°C and 33°C. The FCT has six area councils, namely Abuja Municipal Area Council (AMAC), Bwari, Gwagwalada, Kwali, Kuje and Abaji. AMAC is the area council where the seat of federal government, its agencies and diplomatic offices are located. It has the highest infrastructural development and is residence to politicians, wealthy Nigerians and diplomats. The other area councils are satellite towns with lesser infrastructural development and resident to mostly civil servants, farmers, artisans and traders. The choice of FCT for this study is purposive because it has multi-class consumers of different socio-economic characteristics who have varying demand strength and consumption behaviours. Virtually all imported and local rice brands can be found in the major markets in these area councils.

Federal Capital Territory has a total population of about 3.5 million people (NPC 2013) out of which at least 70% (2.45 million people) are rice consumers who constitute our target population of about 490 000 households (based on average of five people per household). Following the method used by Yamane (1967), this household population gives a sample size of 400 households which is considered adequate for interview and data collection. To cover wider geographical area of the FCT-Abuja, a multi-stage random sampling method was used in selecting a total of 460respondent households as follows: AMAC (76),² Kuje (77), Gwagwalada (77), Abaji (77), Kwali (76) and Bwari (77). Sampling frames were obtained from the Federal Capital Development Authority (FCDA) and Abuja Geographical Information System (AGIS 2014).

Data were collected using a structured and validated questionnaire. Jury’s method was used to test questionnaire content validity (Wallace et al. 2003), while the test–retest method was used to evaluate questionnaire reliability. The questionnaire was primarily administered to the household heads during the face-to-face interview, while other household members contributed in providing answers to the questions raised during the interviews. Data were collected on the consumer households’ socio-economic characteristics, level of perception of the quality attributes, market prices at which they buy and maximum prices they are willing to pay for imported rice brands.

Many studies have used various methods such as choice experiments and contingent (Goldberg & Roosen 2005), stated choice experiment (Travisi & Nijkamp 2004), conjoint analysis (Ara 2003), survey rankings and ratings (Quagrainie 2006), travel cost (Gonzalez & Loomis 2006), experimental auction method (Yue, Alfnes & Jensen 2009) and discrete choice modelling (Haab & McConnell 2002) to estimate WTP as a measure of monetary value of quality attributes of food products as perceived by consumers. Value perception can be defined as a judgement or a valuation by the consumer of the comparison between the benefits or utility obtained from a product, and the perceived sacrifices or costs (Zeithaml 1988). The price consumers are willing to pay for each quality attribute reflects the level of utility derived and the relative importance attached (Sonata & Rasa 2010). The higher the level of a desirable attribute in an alternative food product, the higher the utility associated with that alternative and more likely the consumer is willing to pay a higher price for it (Bennet & Blamey 2001).

Assuming the market price P_ij that ith consumer pays for imported rice j rises to a new price level P_imax, and if the consumer is willing to pay this price increase in order to keep deriving the same level of utility U_ij as previously, the linear utility functions for imported rice at the market price P_ij and at higher price P_imax are expressed in Equation 8 and 9 as follows: Uij=β0j+β1Pij+β2jXi+εij,[Eqn 8] Uij=β0j+β1Pimax+β2jXi+ε2ij,[Eqn 9] β0j+β1Pij+β2jXi+εij=β0j+β1(Pij+WTPij)+β2jXi+ε2ij,[Eqn 10] where P_imax= (P_ij + WTP_ij) and is the maximum price that ith household can pay for imported rice.

As not all of the β_0js and β_2js are identifiable, and Equation 8 and 9 provide the household with the same level of utility, so normalisation rule was adopted such that β_0js and β_2js in Equation 8 become equal to zero (Chern & Chang 2009; Greene 2000). Solving Equation 10 for WTP_ij gives: WTPij=β0j+β2jXi+(εij−ε2ij)β1.[Eqn 11]

Taking the expected value of the WTP_ij, the expected WTP higher price for imported rice in order to remain at the same level of utility and avoid local rice is expressed as: E(WTPij|Xi)=β0j+β2jXiβ1,[Eqn 12] where WTP_ij is ith household WTP for increase in the price of imported rice, X_i is a vector of explanatory variables that influence ith household’s purchase decisions, β_2j is the vector of estimated coefficients of all the explanatory variables and β₁ is the coefficient of P_imax. Following the same procedure in Equation 5, the probability that ith consumer household is willing to pay a higher price P_imax for imported rice j in order to remain at the same level of utility U_ij can be expressed in terms of the logistic distribution (McFadden 1980) as follows: Pr(WTPij=1)=expUij1+expUij,[Eqn 13] where Uij=β0+βiXi and −∞<Uij<+∞.[Eqn 14]

All the explanatory variables X_i remain as previously defined except that the market price P_ij is replaced with P_imax in Equation 14. Following Latvala (2010), the deciding rule is that: if Pr (WTP_ij = 1) > 0.5, there is consumer preference for imported rice brands; and otherwise if Pr (WTP_ij = 1) ≤ 0.5.

The explanatory variables hypothesised to explain consumers’ preference and WTP premium price for imported rice brands were identified based on the theoretical framework and on past empirical work on consumer behaviour towards food products (Bonti-Ankomah & Yiridoe 2006; Millock et al. 2002; Zeng & Wei 2005). The explanatory variables are classified into three categories: price, household socio-economic characteristics and the strength of consumer’s general perception of the quality attributes of imported rice such as neatness, duration of cooking, after-cook colour, aroma, taste, grain shape, swelling capacity, stickiness and texture. The definitions for the variables used in the analysis are presented in Table 1.

The socio-economic characteristics of the respondent households are presented in Table 2. About 65% of the respondents are female; this is not unexpected, because women take most purchasing and dietary decisions and generally tend to have more knowledge and bargaining power for food items as compared to men. The average age of household heads is 47 years, while the average number of years spent in formal schooling is 16 years, indicating that household heads are relatively young and educated, with an average household size of five people. Younger and educated households are more likely to prefer and adopt urban lifestyles and food items that are easy to prepare (such as imported rice) than older and less-educated households. The average monthly income of a household head is ₦88 350 (about $441), indicating that households live on an average of about $10per day, which is well above the national monthly minimum wage of ₦18 000 (about $90). This is because most household heads in the study area are senior private and public employees of government, corporate organisations, international agencies, politicians and top businessmen earning higher wages and income.

Generally, consumers in Nigeria classify rice as ‘local’ and ‘imported’. The distribution of households by their preference and WTP for imported rice brands in the study area is presented in Table 3.

In all the six study areas, 95% of the consumer households expressed preference for imported rice as the brand of choice, while 52% of the households are willing to pay higher price for imported rice brands. Preference for imported rice brands is highest (97%) in AMAC, Kuje and Gwagwalada, which are locations nearest to the city centre, Abuja. While 93% and 94% of respondents in Abaji and Bwari, respectively, also indicated preferences for imported rice. WTP is highest among 50% of households in Kuje but lowest (38%) among households in AMAC. The above descriptive analysis shows that people residing in urban areas with high infrastructural development, population density, income level and economic activities are more likely to express higher preference for food products they perceived to possess better quality attributes such as neatness, easy to prepare, lower cooking duration, and so on. This is also an indication that the preference and WTP for imported rice brands is higher in urban areas and cities where the consumers have the ability to pay higher prices for food items they perceive as being of superior quality. Given the higher population growth in semi-urban and rural areas, it is likely that consumer inertia against preference for imported rice brands could be broken in the future when the population of consumers in cities and urban areas becomes lower than the population of consumers in semi-urban and rural areas.

The results of the two separate binary logistic regression models estimated for consumer’s preference and WTP for imported rice brands are presented in Table 4.

Both models gave correct predictions of 80% and 96% of households’ preference and WTP for imported rice, respectively. Also, in the two models estimated, the Nagelkerke’s R²are 0.56 and 0.62; the Hosmer and Lemeshow (H-L) tests show significance values greater than 0.05, while the chi-square tests of 2 Log Likelihood are significant at 1%. These indicate that there is no significant difference between observed and model-predicted values, implying a moderately strong relationship between the predictors and the prediction. Therefore, the two estimated binary models provide quite good fits and strong explanatory power. In this study, there is absence of multi-collinearity in the two estimated models because the variance inflation factors (VIFs) for all the variables included in the two models were less than 10 (Menard 1995).

The coefficients of parameter estimates of the binary logit model only provide the direction of the effect of the independent variables on the dependent (response) variable and do not represent the actual magnitude of change or probabilities. Therefore, the marginal effects from the binary model, which measures the expected change in probability of a choice being made with respect to a unit change in the independent variable, are reported as the exp (β ) in Table 4. Estimated coefficients for households’ choice of imported rice brands are compared with local rice as the base reference choice.

The estimated coefficient for household head’s age is positively and statistically significant for the probability of higher household WTP prices for imported rice brands, implying that an increase in the age of household heads is more likely to influence the household to choose imported rice brands over local rice. The marginal effects suggest that a year increase in the age of the household head is likely to increase his or her choice of imported rice brands by 9.9% relative to local rice. The reason could be that the older the household head is, the more likely he or she has over the years developed stronger negative perception on poor-quality attributes of local rice, and has built his or her taste around imported rice brands. This is consistent with the findings of Akaeze (2010) that consumption of imported rice brands is more of a mindset and habitual persistence. Campbell et al. (2009) suggested that consumer’s negative attitude (a mindset) against local rice is stronger among older people who, in the past 1–2 decades, have had bad experience such as the presence of chaffs, impurities, uneven grains, long cooking duration and so on, in consuming local rice.

The estimated coefficient for household income is positive and statistically significant for the probability of household’s WTP higher prices for imported rice brands. This implies that an increase in household income increases the probability of household’s WTP for imported rice brands. However, the marginal effect suggests that an increase in income is not likely to influence a household’s WTP for imported rice brands over the local rice. A possible explanation is that a household may prefer imported rice to local rice but may not be willing to spend additional portion of her income to pay for any increase in the price of imported rice brands. This may indicate increasing consumers’ acceptability and competitiveness of local rice in the market. This is consistent in part with a recent study by Alfred and Adekayode (2014) who found that a large percentage of Nigerians consume local rice. Some of the socio-economic variables describing the respondents, such as gender, household size, education and marital status included in the binary logistic model, were not statistically significant for influencing consumer’s preference and WTP higher prices for imported rice brands. According to Enneking (2004), this non-significance of socio-economic characteristics is a typical phenomenon in studies focused on consumer choice.

The estimated coefficient for price is positive and statistically significant for the probability of household’s preference and WTP for higher price for imported rice brands. This implies that an increase in the price of imported rice is likely to increase the probability of household’s preference and WTP a higher price for imported rice. This is because higher prices often lead to higher quality perceptions (Raghubir & Corfman 1999). The marginal effect suggests that an increase in market price increases the odds of surveyed household’s preference and WTP higher price for imported rice brands over the local rice by 0.001% and 0.05% respectively. A possible reason could be household’s belief that increased prices of imported rice brands reflect improved quality, and this belief is being reinforced by the perception that imported rice brands have always had better quality attributes than local rice brands. This supports previous finding of Shiraia (2015) whose appeal leads to favourable price perceptions and purchase intentions when the product price is high. Also, previous report by Campbell et al. (2009) shows that consumers tend to set minimum-quality standards and are unlikely to shift from imported rice to local rice just because the price of the former has risen, instead they tend to shift to lower quality but cheaper imported rice.

The estimated coefficient for consumer household’s general perception is positive and statistically significant for the probability of household’s preference and WTP for imported rice brands. This implies that an increase in the household’s general perception of quality of imported rice is likely to increase the probability of household’s preference and WTP for imported rice. The marginal effect suggests that an increase in household’s general perception from moderate to strong increases the odds of surveyed household’s preference and WTP for imported rice brands over the local rice by 188.9% and 134.5%, respectively. This indicates that consumer’s perception of the quality attributes of food product could be the highest determinant of consumer’s preference and WTP (Akaeze 2010; Thanasuta & Metharom 2015). This could be because households’ general perception reflects the total importance attached to the quality attributes of the food product (Chiliya, Herbst & Roberts-Lombard 2009; Raghubir & Corfman 1999).

In this study, consumers’ preferences and WTP increased prices for imported rice brands were determined by estimating their respective probabilities relative to local rice and the results across the six locations surveyed are as shown in Table 5.

Table 5 shows that probabilities of household’s preference and WTP for imported rice brands vary across the six locations surveyed. The results show that, on average, the probability that a household prefers imported rice brands over the local rice is 95.6%, while the probability that a household does not prefer imported rice brands over the local rice is 4.6%. Similarly, the probability that a household is willing to pay increased price for imported rice brands in order to avoid the local rice is 75.8%, while the probability that a household is not willing to pay increased price for imported rice brands in order to avoid the local rice is 24.2%. The overall implication is that, while many households may prefer imported rice brands over the local rice, few may actually be willing to pay increased price on imported rice brands in order to avoid local rice. This is consistent with the theory of demand in which higher price leads to lower demand. It also agrees with a recent study by Alfred and Adekayode (2014) whose findings show that large percentage of Nigerians were indifferent in their preference of local and imported rice brands. This is an indication that consumers are beginning to accept local rice as a perfect substitute for imported rice brands.

In this study, consumers’ inertia against preference and WTP for imported rice brands is measured by the difference between price and quality differentials of local and imported rice brands as presented in Table 6. The marginal willingness to pay (MWTP) is a measure of the monetary value of a household’s WTP based on its perception of the difference in the quality attributes of local and imported rice brands.

The results in Table 6 revealed that, with higher negative price–quality differential gaps, consumers’ inertia against preference and WTP for imported rice brands is stronger among households residing in the more developed locations (AMAC, Gwagwalada and Kuje) with proximity to Abuja, the FCT city centre. Consumer households residing in locations such as Bwari, Kwali and Abaji that are further away from the city centre and with lower economic activities seem not to see the market price differential as a true reflection of quality differential between local and imported rice brands. Generally, there is negative price–quality differential gap between local and imported rice brands, and this could be a possible reason for the persistence of consumers’ inertia against preference and WTP for imported rice brands in Nigeria. This differential gap of ₦396 (US$2) per 50kg bag that is lower than ₦2000 (US$15) found by Erestein et al. (2003) indicates a possible growth in consumers’ acceptability and competitiveness of local rice in Nigeria as earlier found by Alfred and Adekayode (2014).