Purpose - The purpose of this paper uses household survey data to model the correlates of household consumption and poverty in Fiji. A multivariate empirical analysis is conducted to ascertain those household and community characteristics that correlate with household welfare and poverty. In particular, the results will show how a particular characteristic will affect household poverty conditional on the level of other characteristics that are also potential determinants of poverty. Design/methodology/approach - The key approach in this paper is similar to Mukherjee and Benson and is based on ordinary least squares (OLS) modeling of the natural logarithm of total per capita consumption of households, which serves as the household welfare indicator, against a set of exogenous determinants such as household and community characteristics. For robustness checks, a probit regression is also estimated with the probability of a household being in poverty as the dependent variable and an identical set of independent variables used in the OLS regression. Findings - The results show that higher levels of education, supporting agricultural growth policies in rural areas and reallocation of labour into the formal sector of the economy will prove effective in reducing poverty at the household level. These results can have important policy implications for design and implementation of poverty reduction policies. Originality/value - Empirical studies on the determinants of household consumption and poverty are non-existent in Fiji. This thus is the first study which attempts to model the determinants of poverty at the household level and is the major contribution of this paper.