Using econometric modelling to predict demand for fluid and farm milk: A case study from Turkey

A censored system of demand analysis to unpacked and prepackaged milk consumption in Turkey

“…It could further be argued that this group at higher income levels would tend to curtail the consumption of unpacked milk and simultaneously substitutes more packed milk with increased income as the household regards the consumption of unpacked milk as "inferior good". The results were consistent with previous findings obtained from analysis of household demand patterns for different milk type products (Ratnam and Spielmann 1972;Gould et al 1990;Cornick et al 1994;Gould 1996;Yen et al 2004;Akbay et al 2007;Armagan and Akbay 2008;Pazarlioglu et al 2007). …”

“…Households with kids aged 0-1 and 2-16 demand significantly less on unpacked while more on packed milk products than households without kids in that range. The results have intuitive implications: households with kids demand more the packed milk, whilst they reduce their unpacked milk consumption, possibly because of increased health concern about daily dietary calcium intake along with hygienic (Cornick et al 1994;Akbay et al 2007;Pazarlioglu et al 2007;Armagan and Akbay 2008;Akbay and Tiryaki 2008). Parenting triggers focusing on nutrition search for nurturing benefits via provisions of milk may solidify the health benefits for kids (Gilbert 1997(Gilbert , 2000Verbeke 2005).…”

“…Heterogeneity in preferences in such surveys is accommodated in applied micro-economic analysis through collecting detailed demographic characteristics of household or individual (Yen et al 2004). Among studies based on survey data, Hatirli et al (2004) focus on main factors that affect fluid milk purchasing sources in Antalya province using a probabilistic model (e.g., multinomial logit) while Pazarlioglu et al (2007) analyse the demand for both street milk (e.g., farm milk) and fluid milk in single province, Izmir, using Heckman sample selection model. Akbay et al (2007) use household data from the 2003 Turkish Household Expenditure Survey and focus on aggregated food consumption patterns using a two-step system.…”

“…The procedures that do not specifically account for censoring in the dependent variables result in biased and inconsistent demand estimates (Yen et al 2004;Yen 2005a,b). Pazarlioglu et al (2007) and Akbay et al (2007) did realize to accommodate censoring with the two-step procedure of Heien and Wessells (1990) but did not correct the standard errors of the parameter estimates at the second step. Shonkwiler and Yen (1999) showed that the aforementioned procedure is incorrect analytically and performs poorly in Monte Carlo simulations even in the presence of corrected standard errors of parameters of the second step.…”

“…First, this paper quantifies factors affecting both the quantity demanded of street and packed milk consumptions using a full information maximum likelihood of the bivariate Tobit model. Second, all previous studies with the exception of Pazarlioglu et al (2007) who only partially estimated milk demand elasticities associated with price variables ignored key policy tools such as own-and cross-price, income and relevant variable elasticities in their analysis. We, therefore, obtain expected conditional and unconditional mean levels of both street and packed milk quantities demanded and related unconditional elasticity estimates to quantify the percentage impact of covariates on the quantity demanded of each milk product.…”