Complete markets imply the separation of food production and consumption decisions such that they can be modeled to occur sequentially and can be studied independently. Separation is very often assumed implicitly in empirical studies of food demand. If there is such separation, then food sourced within the household should not have any influence upon the budget share of each food group. Using this insight, this paper first develops a procedure to test for the separation of household food production and consumption decisions. Furthermore, it incorporates the testing procedure into the Almost Ideal Demand Systems model and utilizes survey data from 2003 for Turkey for empirical testing. It concludes that the separation assumption is unwarranted for Turkey. Next, It investigates the extent of bias in elasticity estimates when separation assumption is unwarranted. It concludes that ignoring the nonseparation of consumption and production decisions in rural areas leads to significant overestimation of food expenditure elasticity for the dairy products and eggs and own-price elasticity for bread and cereals.