Analysis of differences of the consumption of mobile voice service based on quantile regression model

This paper researches the demand function of the per capita consumption of mobile voice service by fixed effect quantile regression model. Results show that both price elasticity and income elasticity vary between different consumption-level consumer groups. For low-consumption consumers, income is the most important factor that impacts their consumption, while the price of mobile voice service and the price of communication tools and local fixed voice service all have little effects on consumption; for middle and high consumption consumers, when consumption rise, the price sensitivity of consumers tends to increases, cross-price elasticity of communication tools and local fixed voice service also increase. These results indicate that quantile regression model can give us more detailed and accurate descriptions of demand function than ordinary regression models, and provide us a new method to correctly judge the development stage of the mobile voice service market and then further develop effective marketing strategies.