the own termination rates (see e.g. Armstrong (2002)). As termination rates are costs
for competitors one should expect investments also to affect competitors’ calls to other
networks (see Valletti and Cambini (2005)). Second, the cost-reduction changes the
investor’s amount of outgoing traffic because lower retail prices induce customers to in-
crease their demand for outgoing calls.
I consider these hypotheses under alternative pricing schemes (linear and non-linear)
and compare the results assuming alternative forms of regulation which are in place in
European countries. Afterwards the theoretical findings will be analyzed by adopting
data for the EU 15 countries as well as for Norway and Switzerland. I keep the mod-
els as close as possible to the standard theoretical literature (among others Armstrong
(1998), Laffont, Rey and Tirole (1998a and b), Carter and Wright (2003)) and adopt the
assumptions and approaches provided. This way, comparisons of the empirical outcomes
with the theoretical findings are facilitated and, moreover, also comparisons with the
results of investment effects expected from the literature are allowed for.
In the theoretical part I assume a three-step model with asymmetric players which mainly
corresponds to the model in Dewenter and Haucap (2005).2 In doing so a positive invest-
ment effect on own profits is found but also a positive externality on competitors’ off-net
profits from incoming calls. While non-linear pricing provides similar results as linear
pricing, regulation (as assumed in the literature) ignores network externalities leading
to a deterrence of competitors’ prices and traffic.
With the empirical model I find support that investments reduce the investor’s termi-
nation rate and increase the investor’s incoming traffic. As the traffic effect outweighs
the effect on termination rates in the investor’s short-run profit function the empirical
results support the results expected from the theoretical model. Moreover, investments
increase competitors’ incoming traffic and reduce their termination rates. Replacing the
regulation control variables by interaction terms with investments shows that the neg-
ative effect on competitors’ termination rates is not due to regulation. Combining the
empirical findings with the theoretical model the (pure) investment-induced termination
rate reduction even outweighs the price-driven traffic increase in the competitors’ profit
functions.
2 I rely on the central assumptions in this paper as, to my knowledge, it is the first paper which analyzes
an issue in mobile network competition from both a theoretical and an empirical perspective.