Products demand and substitution modelling and estimation for microdata

Let us consider wide and deep assortment, which is typical for a multiproduct producer or for a wholesaler. Commodity price affects its an other commodities quantities saled to the customer. We mean price demand relations, that should be taken into account by the ologopolist that prices his commodities using market formula. Therefore we propose our theoretical model of substitution that holds own price effect, price substitution effect and income effect. We define price as discrete variable with low, middle and high range of values and model theoretically demand for two substitutes concerning their prices relation. The model is appropriate for count sales data of individual customers. We use ANOVA method for individual demand analysis in groups of customers. Customers are selected by random stratified sampling.
Our model is verified with 2-factor ANOVA of main and price interaction effects. Feasible method of our model verification is also Poisson regression, which we estimate in the paper. Results of both estimation methods are coherent and approve our theoretical model of demand relations between substitutes