Multinomial Logit Model Application: Usage In Beef Market Choice Analysis In Kampala, Uganda

Beef markets in Uganda comprises of supermarkets, abattoirs and butcheries with the latter constituting the biggest share in beef supply and demand. Each market alternative can be identified with distinct attributes of beef quality. Distinct socio-economic categories of consumers and characteristics exist. These may foster or constrain consumers’ utility associated with buying beef from a particular market and hence their preference and market choice, as rooted in the probabilistic and utility maximization theories. Understanding consumers’ underlying market choice determinants is quite relevant to guide policy and strategic interventions to develop beef markets and enhance performance of the beef industry. This studyutilized descriptive statistics and statistical modeling analytical approaches to determine the factors that influence beef consumption and choice between alternative beef markets among beef consumers in urban household of Uganda. Through a face-to-face interview, data from 300 beef consumers were collected. The analysis entailed utilization ofcross tabulations, ANOVA models and more intensively the Multinomial Logit Model. For purposes of hypothesis testing and drawing inferences, Chi-square and T-test statistics were employed. Descriptive statistics indicated that the majority of households consume beef on a weekly basis and at an average of 3.8 kg per week. The ANOVA results revealed that income, education level and house hold size significantly (P<0.05) influence beef consumption among urban households in Uganda. T-test results indicated that with no income constraint, households would significantly increase beef consumption by an average of 0.5Kg per week. Beef consumption was significantly higher among households with more members, earning more income and with higher education level.Likewise, education level had a similar nature of effect on beef consumption.The distribution of the Cox & Snell R Square and the Nagelkerke R2 values suggested that the fitted model with socio-economic variables accounted for 9.5-13.4% percent of the variation in consumers’ choice among alternative beef markets. The probability distribution of the final chi-square for the log likelihood ratio was less than 5% significance level for the overall model and independent variables income and education level but greater than for sex and household size. Thus, income and education level can be used while sex and household size cannot be used to distinguish or characterize consumers who opt for a particular type of market for beef. The logit model estimates indicted that increase in income or education level significantly increasethe likeliness of buying beef in the supermarkets relative to butcheries and supermarkets relative to abattoirs. Increase in education level furtherincrease the likeliness of buying beef from the abattoirs relative to butcheries.The likeliness of buying beef in supermarkets than butcheries and supermarkets than abattoirs given increase in income was estimated at 76.5 and 79.2% respectively. While the likeliness that more consumers will buy beef from supermarkets than butcheries and abattoirs than butcheries given at higher education levelsestimated at 78.7, and 62.6% respectively. Higher income and (or) more educated consumers were less likely than those of low socio-economic status, to consider price of beef but more concern about, hygienic conditions and convenience and hence more likely to buy beef in supermarkets. Comparably, the low income earners and the less educated, in a bid to circumvent the high price of beef, transport costs for reaching supermarkets located in far proximities and in addition to their demand for fresh beef, were more likely to buy beef at butcheries. The study provides key insights into strategic interventions by stakeholders to enhance beef market competitiveness.
Usage In Beef Market Choice Analysis In Kampala, Uganda
Date of publication: 
Region Focus: 
East Africa
RUFORUM Theses and Dissertations
Licence conditions: 
Open Access
Access restriction: 
Prof. Kavoi Muendo (JKUAT) and Prof. Pamela Abbott (IPAR, Kigali Rwanda)
Printed resource