“…= ∏ ′ , and ln ( ) We included six demographic variables in the model to control the variation of preference due to the differences in the demographic characteristics of the household as follows: household size; number of children under five years old; urban/rural classification; educational attainment of household head (completed senior high school or not); income groups (40% lowest income, 40% medium income, and 20% highest income); and employment sector of household head (agriculture or nonagriculture). We then simplify the estimation of the demand model by aggregating all food commodities into the following 14 groups: rice, non-rice staple, tubers, fish, meat, eggs, milk, vegetables, pulses, fruits, oil and grease, beverage ingredients, spices, and other foods, following Faharuddin et al (2017).…”