LIBERALIZATION IN SERVICE NETWORKS: The size of the gains 3

rs_03_sanal_ag
To empirically parameterize the model, we need estimates of preference and technology parameters for each of these regions in the model. In the absence of literature estimates on the relevant elasticities, we have used a value of unity for the transformation elasticities in each region, which we then vary in subsequent sensitivity analysis.

The X parameters in preferences determining the strength of the network externality effects are also important, and we have assumed them to be 0.1 in all countries. Using these assumptions, we calibrate share parameters of preference functions and production frontiers using 1991 data on call volumes, implied call prices, and GDP published in the World Telecommunication Development Report (WTDR) (ITU, 1994) and the World Development Report (WDR) (World Bank (1993)).

The 1994 WTDR gives detailed information on the use of telephone networks by country and region, reporting volumes of both international and domestic calls, and revenues generated by the telephone sector in each country, again broken down by domestic and international calls. It also reports total and per capita GDP, as well as other information such as the telephone density and major telephone operating companies in 202 countries and autonomous territories.

We construct base case data on prices and quantities for our model by dividing WTDR revenues by volumes, using the strong assumption of homogeneity of calls within each group for each country. We use country specific information for Canada and the USA, and use summary statistics for Europe and derive data on the Rest of the World as a residual from world totals minus Canada, USA and Europe. In order to complement this data on trade in network related services to produce a benchmark equilibrium data set for use in calibration, we use merchandise trade data published in the World Development Report, 1993.