Abstract

Study area and sampling technique

This study was conducted in Addis Ababa, the capital city of Ethiopia, which serves as a social, economic, and political center for the country. The total population as of August 2019 was estimated to be 4,592,000 with an annual growth rate of 4.4 percent (Central Statistical Agency [CSA] 2019). The city has a higher level of business activities, including furniture businesses, and larger numbers of furniture manufacturing firms are located there than in any other city or town in the country (CSA 2014).

In this study, a multistage sampling technique was employed to achieve the objectives of the study. In the first stage, Addis Ababa city was selected among the major cities in the country. The study city was chosen as the research area for several reasons: (1) Geographically, Addis Ababa has the highest population with various backgrounds of citizens; (2) there is availability of both modern retail outlets and traditional markets; (3) it is a region that holds a good mixture of potential respondents with different levels of education, income distribution, and ethnicity, which are anticipated to have some impact on the purchase of furniture; and (4) a large consumer base is present for both imported and domestic products, including wood products. Furthermore, the city is the largest city among the cities of Ethiopia, and it is the major destination for imported furniture, with most domestically produced furniture products being transacted there.

In the second stage of the study, Bole and Arada subcities were chosen out of 10 subcities in Addis Ababa due to their furniture potential. Arada subcity was selected because it represents the center of the city and has a wide variety of domestic wooden furniture products and consumers. On the other hand, Bole subcity was chosen for its abundance of many imported wooden furniture products, consumers, and the expansion of new houses, particularly condominiums, to better represent the consumer base.

In the third stage of the study, sample consumer households were selected randomly. The target population or sample unit of the study comprised imported and domestic household wooden furniture consumers. The selected sample frame consisted of consumers with knowledge and experience in purchasing domestically produced and imported furniture within the last 5 years, taking into consideration inflation and changes in consumer behavior. The sample consumers were chosen from populations in both Arada and Bole subcities who had purchased either locally produced or imported furniture, were between the ages of 20 and 65 years, and were actively responsible for making purchasing decisions.

However, the target population was very large. Therefore, the sample size determination method was applied to select representative samples from the total population. To determine the sample size out of the total population of wooden furniture consumers in the sample subcities, the researchers followed Cochran’s formula. The sample size was determined using Cochran’s formula for calculation from a large population (Cochran 1977) as follows:

Thus, the sample size is equal to 385 consumer households. Here, n is the sample size, Z is the selecting criteria value of the desired confidence level (1.96), p is the estimated proportion of an attribute that is present in the population (0.5), q = 0.5 (1 − p), and e is the desired level of precision (0.05). The sample frame was drawn from buyers of imported, domestic, and imported plus domestic wooden furniture in the two sample subcities of Addis Ababa. The sample size from each subcity was determined using proportional to population size. As a result, 122 consumers from Bole subcity and 263 consumers from Arada subcity were included in the sample. In total, 174 respondents purchased imported furniture, 151 respondents purchased domestic furniture, and 60 respondents purchased imported plus domestic furniture.

Data collection methods

Data were collected from primary sources. The main primary data sources were imported and domestic wooden household furniture consumers. The primary data included socioeconomic variables, wooden furniture attributes, consumer perception of furniture attributes, and types of wooden household furniture purchased from imported and domestically produced furniture shops, such as bedroom and living room furniture. The data included the preference of bedroom and living room furniture attributes such as durability, finishing, design, color, price affordability, user-friendliness, compatibility with other objects, guarantee, and multipurpose.

The data were collected from consumers of imported and domestic wooden furniture using structured questionnaires through face-to-face interviews at homes and furniture stores. A self-administered questionnaire was used for data collection. The responses of each participant were recorded on the structured questionnaire. Consumers were categorized into three groups: those who purchased imported furniture only, domestic furniture only, and a combination of imported and domestic furniture. All sample consumers were aware of the manufacturing sources of their purchased furniture. Their preferences were determined by asking why they preferred the specified sources or country of origin for the three furniture groups. During data collection, data were gathered from a total of 390 respondents. Out of these, only 385 responses were deemed valid, while the remaining five responses were found to be invalid and were therefore removed from the analysis.

Based on Fisher’s view of a structured questionnaire, closed-ended questions were employed. If the researcher wants to quantify the research material, it is best to use a structured approach (Fisher and To 2012). The relevant attributes of wooden furniture were identified from various stakeholders (producers, consumers, traders, and SME offices) through key informant interviews. In total, 20 key informants were interviewed to inquire about preferred furniture attributes and the availability of imported and domestic furniture. Prior to actual data collection, a preliminary survey was conducted, and, finally, primary data collection was carried out using the final questionnaire.

Method of data analysis

Multivariate probit model specification and analysis.—

To analyze the major factors influencing consumers’ preference for imported and domestic wooden furniture products, the multivariate probit model (MVP) was utilized. The MVP model is a type of random utility model that examines consumers’ choices and preferences. This model is suitable for analyzing discrete choice decisions, such as selecting furniture attributes and the product itself. It functions as an indirect utility function, where individuals with specific characteristics associate an average utility level with the origin of the product choice set (Lancaster 1966). Multinomial models are typically used when individuals can select only one outcome from a set of mutually exclusive and collectively exhaustive alternatives. However, the MVP model is more appropriate for estimating several correlated binary outcomes simultaneously. It can effectively capture the influence of various explanatory variables on different furniture sources. This model takes into account potentially correlated unobserved disturbances and relationships between choices of different categorical dependent variables (Cappellari and Jenkins 2003). In this study, consumers’ choices of products and attributes were not mutually exclusive. There was a possibility of simultaneous selection of furniture sources and potential correlations among these products. Therefore, the multinomial logit model was deemed to be unsuitable for estimating non-mutually exclusive and nonhomogeneous alternatives, and the MVP was instead used to analyze consumer preferences for different products and to estimate multiple correlated binary outcomes simultaneously (Cappellari and Jenkins 2003).

It is assumed that for a given consumer i making decisions considering nonexclusive alternatives that constitute the preference set k of furniture products, the choice sets may differ according to the decision-maker for maximizing his/her utility. Let U₀ represent the benefits to the consumer who chooses domestic wooden furniture, and let U_k represent the benefits of the consumer choosing the k furniture products, where k denotes preference of imported furniture (Y₁), domestic furniture (Y₂), and both imported and domestic (Y₃). The consumer decides to choose the kth furniture product if Y*_ik = U_k − U₀ > 0. The net benefit Y*_ik that the consumer derives from choosing furniture is a latent variable determined by the observed explanatory variable (X_i) and the error term (ε_i):

Using the indicator function, the unobserved preferences in the above equation translate into the observed binary outcome equation for each choice as follows:

Here, Y_i* is an unobservable latent variable denoting the probability of choosing k type of furniture, for i presents a consumer choosing imported furniture/domestic furniture, both imported and domestic furniture.

In the multivariate model, where the choice of several furniture products is possible, the error terms jointly follow a multivariate normal distribution (MVP) with zero conditional mean and variance normalized to unity (for identification of the parameters), where (μ_x1, μ_x2, μ_x3) MVP ∼ (0, Ω), and the symmetric covariance matrix Ω is given by:

Of particular interest are the off-diagonal elements in the covariance matrix, which represent the unobserved correlation between the stochastic components of the different types of furniture products. This assumption means that Equation 4 generates an MVP model that jointly represents a decision to choose a particular furniture product. This specification with nonzero off-diagonal elements allows for correlation across error terms of several latent equations, which represent the unobserved characteristics that affect the choice of alternative products. Different factors that influenced the preference of consumers for domestic and imported wooden furniture are presented in Table 1.

Table 1. Data summary and hypotheses for explanatory variables.

View Fullscreen

Following the formula used by Cappellari and Jenkins (2003), the log-likelihood function associated with a sample outcome is then given by: where ω is an optional weight for observation i, and Φi is the multivariate standard normal distribution with arguments μi and Ω, where μi can be denoted as:

A categorical dependent variable was measured by the probability of consumers purchasing imported, domestic, or both imported and domestic wooden furniture. It was represented in the model as Y₁ for consumers who choose to purchase imported furniture, Y₂ for consumers who choose domestic wooden furniture, and Y₃ for consumers who choose a combination of both imported and domestic wooden furniture (Table 1). The choices of the outcome variables were presented by (0 and 1), which means that if the consumer chooses the product, it is given a value of 1, and it is 0 otherwise. The expected directional hypotheses of each explanatory/independent variable included in the model are presented in Table 1.

Socioeconomic characteristics of consumers correlates with furniture preference

The different socioeconomic characteristics of respondents were summarized using summary statistics (Table 2). During data collection, 390 respondents were solicited, and only 385 responded and were valid. The findings revealed that 52.7 percent of the consumer households were females, and 65.2 percent of sample consumer households were married. The mean age of consumers was 35.1 years old, and the average years of schooling was 7.7 grade. The result also showed that the average years of staying after marriage was 6.72 years, and the sample respondents had 4.43 average family members per household. The average years of staying in the area was 10.78 years, and 45 percent of sample respondents were self-employed. Regarding income, the average monthly income of the respondent was ETB 12,874.94. Furthermore, the results revealed that 47 percent of respondents lived in a condominium house, and 52 percent of them had their own house. During furniture purchasing time, consumers traveled an average distance of 6.84 km from their residence to the furniture purchasing center. The advertising access showed that 83 percent of respondents had advertising access when they purchased a particular type of wooden furniture.

Table 2. Summary statistics of consumers’ socioeconomic characteristics (see Table 1 for definitions of variables).

View Fullscreen

Motives of consumer preference towards imported and domestic wooden furniture

The results showed that 45.2 percent of respondents preferred to buy imported furniture, while 39.2 percent preferred domestically produced wooden furniture. Only 15.6 percent of respondents preferred to buy both imported and domestic furniture (Fig. 1). This finding is supported by the study conducted by Khattak and Shah (2011), which found that consumers living in big cities consider imported products to be of higher value and feel pride in consuming them. Additionally, this finding aligns with findings showing that consumers in developing countries such as Ethiopia are more likely to favor global products over local ones because they perceive global products as superior to domestic ones (Lumintang et al. 2018).

View Figure

• Download as Powerpoint
• Download Figure

Consumers were asked to provide reasons for their preference between imported and domestic wooden furniture, as shown in Figure 2. The results indicated that consumers favored imported wooden furniture products primarily because of their finishing. Additionally, the results revealed that color was another attractive attribute of imported furniture, followed by the availability of various imported furniture options in the market, compatibility with other objects/items, design, and fast delivery to consumers. On the other hand, attributes such as price affordability, durability, guarantee, ease of maintenance, and environmental friendliness were less favored when describing imported furniture.

View Figure

• Download as Powerpoint
• Download Figure

In the case of domestic wooden furniture, respondents stated that durability was the most important attribute for preferring domestic wooden furniture, followed by being environmentally friendly, easy to maintain, affordability of price, and guarantee. Other attributes such as design, compatibility with other objects, finishing, fast delivery to users, availability of multiple alternatives, and color were less preferred criteria for domestic wooden furniture.

The findings of the study indicate that finishing, color, multiple alternatives, compatibility with other objects, and design are the most preferred product attributes for imported wooden furniture. However, durability, environmental friendliness, ease of maintenance, affordability of price, and guarantee are the most preferred product attributes for domestic wooden furniture. The result is compatible with related literature (Pezeshkpoor 2020, Genc and Merdan 2021, Hilal and Şükrü 2022), which found that furniture selection criteria most often are durability, comfort, functionality, design, cost, environmentally friendly, moveability, and finishing. In addition, this finding is consistent with the study by Chanjarika (2013), who noted that domestic furniture was more durable and had fewer varieties compared to imported furniture, while consumers preferred imported furniture due to its availability of many varieties and designs. In line with this, wooden furniture holds a significant market share among consumers in terms of durability and environmentally friendly features (Guzel 2020, Khojasteh-Khosro et al. 2022). Similarly, the durability, design, and finishing of wooden furniture are highly preferred by consumers (Kizito et al. 2012). In other, developing countries, foreign brands are often preferred over local brands due to their better finishing and design (Lee and Nguyen 2017).

Wooden furniture attribute importance ranking by consumers

The factors that consumers considered when purchasing household furniture were ranked on a five-point preference basis, ranging from not very important to very important (Table 3). The results revealed that in the ranking of furniture attributes by consumer preference for bedroom furniture, durability was the most important attribute with a mean rank of 7.16, followed by finishing with a mean rank of 6.32; design with a mean rank of 5.7; color with a mean rank of 5.12; affordability with a mean rank of 4.58, user-friendliness with a mean rank of 4.28; compatibility with other objects with a mean rank of 4.06; guarantee with a mean rank of 4.4; and multipurpose with a mean rank of 3.74.

Table 3. Mean ranking and preference of consumers for wooden furniture attributes.

View Fullscreen

The value of Kendall’s coefficient of concordance (W) (0.24) was found to be significant at a 1 percent significance level, indicating agreement among respondents in ranking bedroom furniture attributes. The value of W also suggested a notable difference in factors influencing the preference for wooden bedroom furniture. Therefore, it can be concluded that there is a significant difference in consumers’ preference for wooden bedroom furniture attributes. This aligns with findings from previous studies (Oztop et al. 2008, Jaderna 2022, Hitka et al. 2024), which highlighted that buyers consider bedroom furniture purchases as long-term investments, with durability being a key factor in consumer preference. Additionally, design, availability, and price were identified as important factors influencing consumer preferences for bedroom furniture purchases (Hitka et al. 2024).

The results from consumer preferences for living room furniture features revealed that durability was the most preferred attribute with a mean rank of 6.41. This was followed by finishing, with a mean rank of 5.94; design, with a mean rank of 5.73; user-friendliness with a mean rank of 5.71; color with a mean rank of 5.08; affordability with a mean rank of 4.53; and compatibility with other objects with a mean rank of 4.05.

These findings are consistent with previous research indicating that durability was rated as the most important factor by consumers (Oztop et al. 2008). Similarly, other scholars have identified durability, design, and price as key influencing factors for wooden household living room furniture purchase among all consumers (Mohamed and Yi 2008, Hitka et al. 2024). Additional studies have shown that durability, design, finishing, and color are significant attributes influencing consumers’ choices of wooden dining furniture (Wiafe et al. 2014). As human lifestyles evolve, there is a growing trend toward personalized and diversified furniture preferences due to ongoing innovations in furniture attributes (Xiong et al. 2023, Zhou et al. 2023).

Determinants of consumer preference toward imported and domestic wooden furniture products

Table 4 presents the estimation of the MVP model for the factors affecting consumers’ preference for wooden furniture, with a total of 385 valid respondents included. The Wald test is significant at a 1 percent probability level, indicating that the MVP model fits the data reasonably well. Likewise, the model is significant because the null hypothesis, stating that the preference for the three furniture choices is independent, was rejected at a 1 percent significance level. The result of the likelihood ratio test in the model indicates that the null hypothesis of independence between wooden furniture preference decisions is rejected at a 1 percent significance level, and there are significant joint correlations for two estimated coefficients across the equations in the models. The correlation between the probability of choosing domestic or domestic plus imported furniture was significant at a 1 percent level of significance (Table 4). The finding of the estimated correlation implies that coefficients are statistically significantly different from zero in one pair of the three pairs. This suggests that the unobservable factors that increase the probability that a consumer chooses to purchase their furniture from domestic sources reduce the probability of purchasing the same from both imported plus domestic sources. The preference for domestic wooden furniture was found to be a substitution for domestic plus imported, as presented by rho32 (Table 4).

Table 4. MVP estimations for determinants of wooden furniture preference (see Table 1 for definitions of variables).

View Fullscreen