This event has ended. View the official site or create your own event → Check it out
This event has ended. Create your own
View analytic
Sunday, September 27 • 9:32am - 10:05am
Analyzing the Characteristic Determinants of Smartphone Post-Paid Pricing in South Korea 2010-2015

Sign up or log in to save this to your schedule and see who's attending!

The objective of this study is to analyze the determinants of smartphone wireless plan prices in South Korea from 2010 to 2015. This study attempts to measure the monetary effect of the key service characteristics -- data, voice, text, speeds, handset subsidy, contract duration, and other additional features. The two elementary hypotheses of this study: (1) Taken separately, each characteristic variable is significantly related to price (2) The interactive effect of the selected two characteristic variables is significantly related to the level of price. For example, it is hypothesized that the relationship between price and data allowance is greater at higher levels of handset subsidy or voice or text messaging. In terms of regression equations, we expect that the data allowance impact on the price is different when we add a series of interaction terms to the explanatory variables.

The study estimates the hedonic pricing equation using OLS regression in which changes in the price are regressed on changes in the quantities of the characteristics. We first employ ratio transformation techniques to keep the collinearity among the regressors under acceptable range, and propose a baseline equation. We use the baseline equation to experiment with four alternative functional forms -- linear, semi-log, double-log, and Box-Cox. The study carries out a series of diagnostic tests for the regression specifications, and compares the F values, R squared values, the variances of the residuals, VIF statistics, and AICs, which allows us to figure out how well each model fits the observations. The study runs not only semi-annually separate regression but also pooled regression over the observations from all time periods. The pooled regression is based on the assumption of fixed slope coefficients, but it can involve the assumption of random intercepts depending on judgment about the presence of any conspicuous structural change over time. The paper does not perform any weighted regression due to the limitations of subscription data. The interpretation of the estimated coefficients is the major interest in this paper.

The study tracks the smartphone standard post-paid offerings offered by Korea’s 3 major wireless carriers from the 2010 1st to the 2015 1st. This panel dataset contains between around 50~80 observations for each semi-annual period, and an overall total number of around 550 observations. The sample represents at least more than 94% subscribers out of the whole market throughout the periods. The dataset can be divided into three groups: (1) Pricing information, considered as indicators of the monthly payment that is equal to the minimum total cost on a 2 year contract at certain interest rate, divided by 24 -- the number of months. The minimum total cost consists of wireless plan service charges and handset payments. (2) Usage allowance information, including data, voice, text, and downstream speed measured by megabytes, minutes, integers, and kilobits per second, respectively. (3) Handset subsidy information, based on a monthly reduction in the price of a postpaid service plan with contract. For the sake of parsimony, the models choose just one type of device on which handset basket rules are based, so we do not consider different handset models. Transaction level data on retail price and characteristics are compiled from many sources, including agreements, firms’ online databases, newspapers, magazines, KISDI wireless competition reports, etc.

The estimated coefficients will assist in disentangling the causes of wireless service pricing that are still complex. Carriers have increasingly moved to inexpensive voice-text messaging packages and changed their pricing schemes around data treatment. However, the extent to which any subsidy reduction or increased voice limits benefit consumers is an empirical question, because it can be set by carriers in response to demand in data usage. Thus the coefficients will provide insights into how firms choose different means of enforcing data limits in the evolution of the industry.

This study builds on the previous fixed-broadband hedonic works including Prud'homme & Yu (2001), Williams (2008), and Greenstein & McDevitt (2011).



Wook Joon Kim

Researcher, Korea Information Society Development Institute

Sunday September 27, 2015 9:32am - 10:05am
GMUSL - Room 332

Attendees (4)