j πii,j are retail profits from outgoing traffic, j πji,i are wholesale profits from incoming
traffic. Off-net call prices depend on the termination network. While termination rates
are set equally for all incoming calls, retail prices for off-net calls are chosen dependent
on the termination rates of the called network.
Let us assume a three stage game where, first, MNPs decide on investments, afterwards
they choose termination rates ti and then decide on retail off-net prices. Customers
choose their amount of off-net communication dependent on their provider’s retail off-
net price. Using backward induction optimum retail prices and termination rates will be
derived. With the structure assumed here I follow Dewenter and Haucap and ignore the
possible long-run strategy for a termination rate choice. The long-run termination rate
choice is the focal subject of many theoretical papers beginning with the seminal work
of A-LRT and Gans and King (2001) from a time-independent perspective and with a
time-dependent perspective in Hoeffler (2007).
Deriving (1) one gets i’s profit-maximizing off-net price:
a ci + tj
(2)
pij = 2bi + 2
Note that provider i only partially passes through termination rates to its customers.
Replacing prices for off-net calls and deriving the resulting profit function with respect
to termination rates yields:
ci ∑j sj (a - bj cj )
(3)
2+ 2 Pj sj bj
3.2 Investments
As we cannot analyze the equilibrium investment behavior of MNPs in the empirical
part of the paper, I use comparative statics here considering the effect of investments
on the investor’s termination rate, retail prices and off-net traffic and the externality
of investments on competitors’ off-net profits. I concentrate on investments in cost-
reduction ki and assume c0i(ki) < 0. The reasoning behind this assumption is that
cost-reducing/cost-efficiency increasing investments are only implemented if the cost
level taking into account depreciation is reduced.