ity is comparably low investments are used to force competitors out of the market.
Kotakorpi (2006) also finds support for the under-investment problem with downstream
substitutability. Following Kotakorpi network competition eliminates the foreclosure
challenge of the upstream monopoly. Though, the long-run under-investment problem
still remains in place.
Cambini and Valletti (2005) adopt the analysis of investment incentives to the frame-
work of network competition by introducing an investment stage to the standard A-LRT
model. The central assumption is that investments increase quality but do not affect
per-unit costs (as in Foros (2004) and Kotakorpi (2006)). They show that with asym-
metric network size a small firm would benefit from a mark-up of termination rates over
costs whereas the larger competitor would lose. With a lower level of substitutability
between the services offered by the operators both providers would reduce investments
with termination rates above costs. Nevertheless, without regulation competitors would
negotiate a termination rate above per-unit costs which reduces the incentive to invest
in quality increase.
I will keep these central findings from theory in mind when analyzing the impact of
investment spill-overs. As there exists nearly no empirical analysis of the theoretical
findings I try to provide some more insight into the interplay of competitors in mobile
markets by adopting the theoretical findings on the effects of investments on off-net
prices and traffic into an empirical framework.
3 Theoretical Model
In this section I derive a theoretical model for short-run profits where I assume market
shares as given. I start with the more restricted linear pricing model and show how
termination rates and quantities change due to investments in cost-reduction. Then I
compare the results of the linear pricing model with the outcome under two-part tariff
pricing. Finally, termination rates are fixed at a constant level due to regulation (i.e.
either at per-unit costs/at a constant rate above per-unit costs or at a cost-independent
level). I employ a simplified version of the model in Dewenter and Haucap (2005) and
use comparative statics to analyze alternative effects of investments on termination rates
and traffic in terms of total minutes of usage (MOU).