Effect of Income Heterogeneity on Price Response During the Diffusion of Wireless Service in Multiple Markets
40 Pages Posted: 25 Aug 2010
Date Written: March 15, 2010
To model diffusion of new products in a way that is adequate to prescribe prices, we propose a modification of the standard random coefficients logit model that allows for changes in the distribution of the sample population due to differential adoption rates. Essentially, we use our structural model to endogenously derive population “weights” to update the distribution over time. The interaction of observable demographics with the parameters of the utility specification helps the model explain differences in the rate of adoption across markets and time. This, together with the fact that our approach is well suited to deal with the potential problem of price endogeneity, makes our model ideal to prescribe prices. We use quarterly data on several European countries to estimate the diffusion of mobile phone services. Since in this category churn is elevated and penetration exceeds 100%, our approach deals with both replacement purchases and multiple adoptions. We find that our model does an excellent job matching both adoption rates and prices in the data. Income, but not age, turns out to be an important demographic to explain diffusion. In addition, other explanations such as network effects and technological changes result in insignificant coefficients.
Keywords: Structural Models, NEIO, Nonlinear Pricing, Network Effects, Adoption of New Products
JEL Classification: C50, D40, L00, M31
Suggested Citation: Suggested Citation